___ ____ ____ ____ ____(R) /__ / ____/ / ____/ ___/ / /___/ / /___/ Statistics/Data Analysis 1 . set seed 1010101 2 . set matsize 11000 3 . 4 . 5 . ******* 6 . 7 . * This script requires pre-installation of spost2 by J. Scott Long for fits > tat command 8 . * It also requires installation of title2.ado, title3.ado 9 . * 10 . * for the fitstat command to function 11 . ****** 12 . 13 . global User "Robert Alan Yaffee" 14 . * 17 June 2012 15 . 16 . *------- Primary objective: Hypothesis 1 tests for Part 2 of Nottingham heal > th profile 17 . 18 . ********* Hypothesis tests Robert A. Yaffee 10 June 2012 main effect > s and moderator identification approach 19 . * protocol is to test Health profile subscales against potential confounders > including socio-demog vars, 20 . * major neg life events, stresses and hassles, social supports, perceived > health, distance, 21 . * threat and dose to see whether there is a dose-psych response 22 . 23 . * We construct a full model, trimmed model (at .1 level), and then test inte > ractions with dose if there is a 24 . * significant main effect dose psych response 25 . 26 . * Trimming is performed with backward elimination to the .1 level 27 . * 28 . * This approach identifies the moderating variables for our path analysis. > We analyze our 29 . * final models for congruence with regression assumptions with the rdiag > program for a validity assessment 30 . 31 . *---------------------------------------------------------------------------- > ------------------------------ 32 . 33 . *----------- Organization of the program ------------------------------------ > -------------------------------------- 34 . * main effects regresssions are run for all part 1 health profile subscales > except that of emotional 35 . * reaction for both males and females separately 36 . * these models are trimmed with backward elimination to determine whether th > e main effect of avg cumulative dose 37 . * is significant for that wave 38 . * If the main effect of average cumulative dose for that wave is not statis > tically significant with the 39 . * covariates controlled we do not go further along that path 40 . * If the main effect of average cumulative dose for that wave is statistica > lly significant, we trim the 41 . * model to a p = .1 42 . * If average cumulative dose is not significant, we stop 43 . * If average cumulative dose is significant we perform a hierarchical regre > ssion with interactions between dose 44 . * and other significant main effects 45 . * We proceed to test for mediating effects of those other significant expla > natory variable 46 . 47 . ***-------------------------------------------------------------------------- > ---------------------------------------- 48 . 49 . * Part 2 Nottingham subscales 50 . * 1: hp2work 51 . * 2: hp2hmcare 52 . * 3: hp2probsoc 53 . * 4: hp2pbfhm 54 . * 5: hp2sexlife 55 . * 6: hp2inthob 56 . * 7: hp2vacatn 57 . 58 . local dvwhole HP2work-HP2vacatn 59 . local dv2 HP2work HP2hmcare HP2probsoc HP2pbfhm HP2sxlife HP2inthob HP2vacatn 60 . 61 . 62 . 63 . ************************************* 64 . di "$User" Robert Alan Yaffee 65 . 66 . title "Testing Hypothesis 1 part 2 wave 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Testing Hypothesis 1 part 2 wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:10:40 ***** ******************************************************************************* ******************************************************************************* 67 . 68 . cd /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/H1tests/h1p > t2 /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/h1pt2 69 . use chwide16june2012, clear (Zero for missing on all icdx) 70 . cd /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/H1tests/h1p > t2 /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/h1pt2 71 . 72 . 73 . di "{hline}" ------------------------------------------------------------------------------- 74 . di "{hline}" ------------------------------------------------------------------------------- 75 . title2 "Chunk 1 Hyp 1:radiation dose and Nottingham Health profile subscales" ------------------------------------------------------------------------------- title2: Chunk 1 Hyp 1:radiation dose and Nottingham Health profile subscales Date and time: 18 Jun 2012 18:10:41 Working directory: /Users/robertyaffee/ > Documents/data/research/chwk/phase3/Htests/h1tests/h1pt2 Stata data file: chwide16june2012.dta > has 2373 variables and 703 observations Chunk 1 Hyp 1:radiation dose and Nottingham Health profile subscales ------------------------------------------------------------------------------- 76 . 77 . 78 . // there is substantial intercorrelation among the items warranting a 79 . // multivariate regression model 80 . cap dummies educ 81 . cap order educ1-educ8, after(educ) 82 . 83 . * These variables are substantially correlated 84 . pwcorr HP2work HP2hmcare HP2probsoc HP2pbfhm HP2sxlife HP2inthob HP2vacatn, / > // > obs sig | HP2work HP2hmc~e HP2pro~c HP2pbfhm HP2sxl~e HP2int~b HP2vac~n -------------+--------------------------------------------------------------- HP2work | 1.0000 | | 703 | HP2hmcare | 0.4878 1.0000 | 0.0000 | 703 703 | HP2probsoc | 0.4587 0.5420 1.0000 | 0.0000 0.0000 | 703 703 703 | HP2pbfhm | 0.2832 0.4150 0.4745 1.0000 | 0.0000 0.0000 0.0000 | 703 703 703 703 | HP2sxlife | 0.4968 0.4576 0.5589 0.4192 1.0000 | 0.0000 0.0000 0.0000 0.0000 | 703 703 703 703 703 | HP2inthob | 0.3787 0.4757 0.5956 0.5089 0.5401 1.0000 | 0.0000 0.0000 0.0000 0.0000 0.0000 | 703 703 703 703 703 703 | HP2vacatn | 0.4166 0.4757 0.5956 0.4416 0.5211 0.6840 1.0000 | 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 | 703 703 703 703 703 703 703 | 85 . 86 . cap gen havmilsq = havmil^2 87 . 88 . cap rename Havmil havmil 89 . // controlling for potential confounders 90 . // socio-demographics age gender educ income occp marstat children inc 91 . // distance from accident side 92 . // perceived Chornobyl related health threat to oneself 93 . 94 . local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 95 . local w2bf bf1 bf4 bf6 bf7 bf14 bf15 bf40 96 . local w3bf bf1 bf4 bf2 bf4m bf5m bf7m bf8 bf15m bf17 bf20 bf22 bf29 bf30 bf40 97 . 98 . *---------------------------------------------------------------------------- > --- 99 . * Hypothesis 1 Part 2 wave 1 tests male and female 100 . * endogneous Nottingham pt 2 subscales: HP2work HP2hmcare HP2probsoc HP2pbfhm > /// > * HP2sxlife HP2inthob HP2vacatn 101 . * structure of models 102 . * 1. general models on all Pt 2 subscales with all potential confounders 103 . * 2. trimmed models on all Pt 2 subscales with from all potential confound > ers 104 . * 3. from trimmed models examination of possible moderator variables 105 . * 4. from trimmed models examination of possible mediator variables 106 . * 5. Summary analysis and model evaluation of final models only 107 . * program is divided into 8 chunks one a general model and 1 for each 108 . * endogenous variable 109 . *---------------------------------------------------------------------------- > --- 110 . * Chunk 1 General models for all part 2 of Nottingham Health Profile 111 . 112 . 113 . forvalues j=1/1 { 2. des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 3. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) 114 . 115 . 116 . title " H1: Test of hypothesis 1: Pt 2 wave 1 general models" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** H1: Test of hypothesis 1: Pt 2 wave 1 general models ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:10:41 ***** ******************************************************************************* ******************************************************************************* 117 . forvalues j=1/1 { 2. set more off 3. foreach var in HP2work HP2hmcare HP2probsoc HP2pbfhm HP2sxlife HP2inthob > /// > HP2vacatn { 4. forvalues k=1/2 { 5. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 6. di _skip(4) 7. 118 . title4 "Full main model for `var' for wave= `j' " 8. di _skip(4) 9. title4 "chunk 2 H1 test:Gender= `k' model Wave = `j' for `e(depvar)' " 10. di _skip(4) 11. title4 "Full Nottingham Part 2 `var' subscale models" "wave `j' for gender > ==`k'" 12. 119 . xi: logit `var' age /// > radhlw`j' bf1 bf20 /// > deaw`j' /// > shhlw`j' shjobw`j' shrelaw`j' suprtw`j' /// > if gender==`k', difficult iterate(500) nolog 13. estat class 14. estat gof 15. fitstat 16. } 17. } 18. } ------------------------------------------------------------------------------- Full main model for HP2work for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for hp2vactn ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2work subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 31.46 Prob > chi2 = 0.0002 Log likelihood = -157.14168 Pseudo R2 = 0.0910 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0308396 .0127975 2.41 0.016 .005757 .0559223 radhlw1 | .0011812 .0045458 0.26 0.795 -.0077284 .0100908 bf1 | -.0366332 .0469292 -0.78 0.435 -.1286127 .0553463 bf20 | .0453017 .0444164 1.02 0.308 -.0417528 .1323562 deaw1 | -.0051463 .2892671 -0.02 0.986 -.5720994 .5618067 shhlw1 | .0001023 .0055692 0.02 0.985 -.0108131 .0110176 shjobw1 | .0091313 .0050059 1.82 0.068 -.0006801 .0189427 shrelaw1 | .0025607 .0044585 0.57 0.566 -.0061777 .0112992 suprtw1 | .0026986 .0050591 0.53 0.594 -.0072171 .0126143 _cons | -5.679852 1.827319 -3.11 0.002 -9.261332 -2.098372 ------------------------------------------------------------------------------ Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 3 4 | 7 - | 67 266 | 333 -----------+--------------------------+----------- Total | 70 270 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 4.29% Specificity Pr( -|~D) 98.52% Positive predictive value Pr( D| +) 42.86% Negative predictive value Pr(~D| -) 79.88% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.48% False - rate for true D Pr( -| D) 95.71% False + rate for classified + Pr(~D| +) 57.14% False - rate for classified - Pr( D| -) 20.12% -------------------------------------------------- Correctly classified 79.12% -------------------------------------------------- Logistic model for HP2work, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 339.90 Prob > chi2 = 0.2483 Measures of Fit for logit of HP2work Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -157.142 D(330): 314.283 LR(9): 31.462 Prob > LR: 0.000 McFadden's R2: 0.091 McFadden's Adj R2: 0.033 Maximum Likelihood R2: 0.088 Cragg & Uhler's R2: 0.138 McKelvey and Zavoina's R2: 0.183 Efron's R2: 0.092 Variance of y*: 4.026 Variance of error: 3.290 Count R2: 0.791 Adj Count R2: -0.014 AIC: 0.983 AIC*n: 334.283 BIC: -1609.269 BIC': 20.998 ------------------------------------------------------------------------------- Full main model for HP2work for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2work ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2work subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 47.56 Prob > chi2 = 0.0000 Log likelihood = -182.19002 Pseudo R2 = 0.1154 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0643256 .0124013 5.19 0.000 .0400194 .0886317 radhlw1 | .0047546 .0039694 1.20 0.231 -.0030253 .0125344 bf1 | .0004355 .0252089 0.02 0.986 -.0489731 .049844 bf20 | .0056912 .0216782 0.26 0.793 -.0367972 .0481796 deaw1 | -.1112272 .1659119 -0.67 0.503 -.4364086 .2139542 shhlw1 | .0037005 .0053166 0.70 0.486 -.0067199 .0141209 shjobw1 | .0025496 .0049315 0.52 0.605 -.0071159 .0122152 shrelaw1 | -.0059036 .0043278 -1.36 0.173 -.0143859 .0025787 suprtw1 | .0007605 .0038753 0.20 0.844 -.006835 .008356 _cons | -5.262441 1.017784 -5.17 0.000 -7.257261 -3.26762 ------------------------------------------------------------------------------ Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 24 12 | 36 - | 69 256 | 325 -----------+--------------------------+----------- Total | 93 268 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 25.81% Specificity Pr( -|~D) 95.52% Positive predictive value Pr( D| +) 66.67% Negative predictive value Pr(~D| -) 78.77% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 4.48% False - rate for true D Pr( -| D) 74.19% False + rate for classified + Pr(~D| +) 33.33% False - rate for classified - Pr( D| -) 21.23% -------------------------------------------------- Correctly classified 77.56% -------------------------------------------------- Logistic model for HP2work, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 374.34 Prob > chi2 = 0.1680 Measures of Fit for logit of HP2work Log-Lik Intercept Only: -205.969 Log-Lik Full Model: -182.190 D(351): 364.380 LR(9): 47.557 Prob > LR: 0.000 McFadden's R2: 0.115 McFadden's Adj R2: 0.067 Maximum Likelihood R2: 0.123 Cragg & Uhler's R2: 0.181 McKelvey and Zavoina's R2: 0.198 Efron's R2: 0.141 Variance of y*: 4.100 Variance of error: 3.290 Count R2: 0.776 Adj Count R2: 0.129 AIC: 1.065 AIC*n: 384.380 BIC: -1702.616 BIC': 5.443 ------------------------------------------------------------------------------- Full main model for HP2hmcare for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for HP2work ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2hmcare subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 46.85 Prob > chi2 = 0.0000 Log likelihood = -149.44808 Pseudo R2 = 0.1355 ------------------------------------------------------------------------------ HP2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .052336 .0131829 3.97 0.000 .026498 .0781739 radhlw1 | .0043343 .0047058 0.92 0.357 -.0048889 .0135576 bf1 | -.0149427 .033119 -0.45 0.652 -.0798548 .0499693 bf20 | .0144084 .0298855 0.48 0.630 -.0441661 .0729829 deaw1 | -.1108583 .2961578 -0.37 0.708 -.6913169 .4696004 shhlw1 | .0046663 .0057067 0.82 0.414 -.0065187 .0158513 shjobw1 | .0132629 .0053838 2.46 0.014 .0027109 .023815 shrelaw1 | -.0051278 .0047295 -1.08 0.278 -.0143976 .0041419 suprtw1 | -.0070556 .0059804 -1.18 0.238 -.018777 .0046658 _cons | -5.506222 1.326203 -4.15 0.000 -8.105532 -2.906912 ------------------------------------------------------------------------------ Logistic model for HP2hmcare -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 15 6 | 21 - | 55 264 | 319 -----------+--------------------------+----------- Total | 70 270 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2hmcare != 0 -------------------------------------------------- Sensitivity Pr( +| D) 21.43% Specificity Pr( -|~D) 97.78% Positive predictive value Pr( D| +) 71.43% Negative predictive value Pr(~D| -) 82.76% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.22% False - rate for true D Pr( -| D) 78.57% False + rate for classified + Pr(~D| +) 28.57% False - rate for classified - Pr( D| -) 17.24% -------------------------------------------------- Correctly classified 82.06% -------------------------------------------------- Logistic model for HP2hmcare, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 340.89 Prob > chi2 = 0.2365 Measures of Fit for logit of HP2hmcare Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -149.448 D(330): 298.896 LR(9): 46.850 Prob > LR: 0.000 McFadden's R2: 0.136 McFadden's Adj R2: 0.078 Maximum Likelihood R2: 0.129 Cragg & Uhler's R2: 0.202 McKelvey and Zavoina's R2: 0.237 Efron's R2: 0.150 Variance of y*: 4.310 Variance of error: 3.290 Count R2: 0.821 Adj Count R2: 0.129 AIC: 0.938 AIC*n: 318.896 BIC: -1624.656 BIC': 5.611 ------------------------------------------------------------------------------- Full main model for HP2hmcare for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2hmcare ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2hmcare subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 81.82 Prob > chi2 = 0.0000 Log likelihood = -191.97101 Pseudo R2 = 0.1757 ------------------------------------------------------------------------------ HP2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0920059 .0127833 7.20 0.000 .0669511 .1170607 radhlw1 | .0020253 .0038466 0.53 0.599 -.0055139 .0095645 bf1 | -.00663 .0212107 -0.31 0.755 -.0482022 .0349423 bf20 | -.0001616 .0177132 -0.01 0.993 -.0348789 .0345556 deaw1 | -.084969 .1571368 -0.54 0.589 -.3929515 .2230135 shhlw1 | -.0017213 .0052283 -0.33 0.742 -.0119686 .0085261 shjobw1 | .0107311 .0048545 2.21 0.027 .0012164 .0202457 shrelaw1 | -.0074675 .0042372 -1.76 0.078 -.0157724 .0008373 suprtw1 | -.0008664 .0038272 -0.23 0.821 -.0083676 .0066348 _cons | -5.472133 .9016353 -6.07 0.000 -7.239306 -3.70496 ------------------------------------------------------------------------------ Logistic model for HP2hmcare -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 62 28 | 90 - | 63 208 | 271 -----------+--------------------------+----------- Total | 125 236 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2hmcare != 0 -------------------------------------------------- Sensitivity Pr( +| D) 49.60% Specificity Pr( -|~D) 88.14% Positive predictive value Pr( D| +) 68.89% Negative predictive value Pr(~D| -) 76.75% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 11.86% False - rate for true D Pr( -| D) 50.40% False + rate for classified + Pr(~D| +) 31.11% False - rate for classified - Pr( D| -) 23.25% -------------------------------------------------- Correctly classified 74.79% -------------------------------------------------- Logistic model for HP2hmcare, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 375.75 Prob > chi2 = 0.1555 Measures of Fit for logit of HP2hmcare Log-Lik Intercept Only: -232.881 Log-Lik Full Model: -191.971 D(351): 383.942 LR(9): 81.821 Prob > LR: 0.000 McFadden's R2: 0.176 McFadden's Adj R2: 0.133 Maximum Likelihood R2: 0.203 Cragg & Uhler's R2: 0.280 McKelvey and Zavoina's R2: 0.288 Efron's R2: 0.223 Variance of y*: 4.623 Variance of error: 3.290 Count R2: 0.748 Adj Count R2: 0.272 AIC: 1.119 AIC*n: 403.942 BIC: -1683.054 BIC': -28.821 ------------------------------------------------------------------------------- Full main model for HP2probsoc for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for HP2hmcare ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2probsoc subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 66.20 Prob > chi2 = 0.0000 Log likelihood = -92.051568 Pseudo R2 = 0.2645 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .065801 .0176318 3.73 0.000 .0312434 .1003586 radhlw1 | .004144 .0061988 0.67 0.504 -.0080054 .0162935 bf1 | .0313968 .0422548 0.74 0.457 -.051421 .1142146 bf20 | -.0193584 .0374978 -0.52 0.606 -.0928527 .0541359 deaw1 | -.5581614 .4525925 -1.23 0.217 -1.445226 .3289037 shhlw1 | .0137667 .0070908 1.94 0.052 -.000131 .0276644 shjobw1 | .0236604 .0075625 3.13 0.002 .0088381 .0384826 shrelaw1 | -.0149185 .006252 -2.39 0.017 -.0271722 -.0026648 suprtw1 | .0010509 .0069016 0.15 0.879 -.012476 .0145778 _cons | -7.048599 1.70409 -4.14 0.000 -10.38855 -3.708644 ------------------------------------------------------------------------------ Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 13 0 | 13 - | 28 299 | 327 -----------+--------------------------+----------- Total | 41 299 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 31.71% Specificity Pr( -|~D) 100.00% Positive predictive value Pr( D| +) 100.00% Negative predictive value Pr(~D| -) 91.44% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.00% False - rate for true D Pr( -| D) 68.29% False + rate for classified + Pr(~D| +) 0.00% False - rate for classified - Pr( D| -) 8.56% -------------------------------------------------- Correctly classified 91.76% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 450.80 Prob > chi2 = 0.0000 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -125.152 Log-Lik Full Model: -92.052 D(330): 184.103 LR(9): 66.202 Prob > LR: 0.000 McFadden's R2: 0.264 McFadden's Adj R2: 0.185 Maximum Likelihood R2: 0.177 Cragg & Uhler's R2: 0.340 McKelvey and Zavoina's R2: 0.458 Efron's R2: 0.278 Variance of y*: 6.073 Variance of error: 3.290 Count R2: 0.918 Adj Count R2: 0.317 AIC: 0.600 AIC*n: 204.103 BIC: -1739.449 BIC': -13.741 ------------------------------------------------------------------------------- Full main model for HP2probsoc for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2probsoc ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2probsoc subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 102.23 Prob > chi2 = 0.0000 Log likelihood = -131.99945 Pseudo R2 = 0.2791 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1172627 .0171673 6.83 0.000 .0836153 .1509101 radhlw1 | .0111397 .0049416 2.25 0.024 .0014543 .0208251 bf1 | .0362951 .0284738 1.27 0.202 -.0195124 .0921027 bf20 | -.0253268 .0231741 -1.09 0.274 -.0707472 .0200936 deaw1 | -.0870921 .1767587 -0.49 0.622 -.4335327 .2593485 shhlw1 | .0025727 .0062142 0.41 0.679 -.009607 .0147524 shjobw1 | .0087888 .0058839 1.49 0.135 -.0027435 .020321 shrelaw1 | -.0093887 .0051896 -1.81 0.070 -.0195601 .0007827 suprtw1 | .0013382 .0045531 0.29 0.769 -.0075856 .0102621 _cons | -8.243948 1.264801 -6.52 0.000 -10.72291 -5.764984 ------------------------------------------------------------------------------ Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 33 14 | 47 - | 41 273 | 314 -----------+--------------------------+----------- Total | 74 287 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 44.59% Specificity Pr( -|~D) 95.12% Positive predictive value Pr( D| +) 70.21% Negative predictive value Pr(~D| -) 86.94% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 4.88% False - rate for true D Pr( -| D) 55.41% False + rate for classified + Pr(~D| +) 29.79% False - rate for classified - Pr( D| -) 13.06% -------------------------------------------------- Correctly classified 84.76% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 399.18 Prob > chi2 = 0.0329 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -183.113 Log-Lik Full Model: -131.999 D(351): 263.999 LR(9): 102.227 Prob > LR: 0.000 McFadden's R2: 0.279 McFadden's Adj R2: 0.225 Maximum Likelihood R2: 0.247 Cragg & Uhler's R2: 0.387 McKelvey and Zavoina's R2: 0.459 Efron's R2: 0.306 Variance of y*: 6.086 Variance of error: 3.290 Count R2: 0.848 Adj Count R2: 0.257 AIC: 0.787 AIC*n: 283.999 BIC: -1802.997 BIC': -49.227 ------------------------------------------------------------------------------- Full main model for HP2pbfhm for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for HP2probsoc ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2pbfhm subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 22.29 Prob > chi2 = 0.0080 Log likelihood = -70.363535 Pseudo R2 = 0.1367 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0531285 .0215313 2.47 0.014 .0109279 .0953291 radhlw1 | .0202681 .008252 2.46 0.014 .0040946 .0364417 bf1 | -.0087893 .0546645 -0.16 0.872 -.1159298 .0983511 bf20 | .007857 .0497616 0.16 0.875 -.0896739 .1053879 deaw1 | -.0860305 .4565197 -0.19 0.851 -.9807926 .8087316 shhlw1 | .005778 .0099994 0.58 0.563 -.0138205 .0253765 shjobw1 | -.0120291 .0099428 -1.21 0.226 -.0315167 .0074584 shrelaw1 | .0007624 .0073877 0.10 0.918 -.0137173 .015242 suprtw1 | -.018106 .0125821 -1.44 0.150 -.0427664 .0065544 _cons | -6.445038 2.186355 -2.95 0.003 -10.73022 -2.159861 ------------------------------------------------------------------------------ Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 0 0 | 0 - | 22 318 | 340 -----------+--------------------------+----------- Total | 22 318 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 0.00% Specificity Pr( -|~D) 100.00% Positive predictive value Pr( D| +) .% Negative predictive value Pr(~D| -) 93.53% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.00% False - rate for true D Pr( -| D) 100.00% False + rate for classified + Pr(~D| +) .% False - rate for classified - Pr( D| -) 6.47% -------------------------------------------------- Correctly classified 93.53% -------------------------------------------------- Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 349.26 Prob > chi2 = 0.1510 Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -81.506 Log-Lik Full Model: -70.364 D(330): 140.727 LR(9): 22.285 Prob > LR: 0.008 McFadden's R2: 0.137 McFadden's Adj R2: 0.014 Maximum Likelihood R2: 0.063 Cragg & Uhler's R2: 0.167 McKelvey and Zavoina's R2: 0.285 Efron's R2: 0.087 Variance of y*: 4.600 Variance of error: 3.290 Count R2: 0.935 Adj Count R2: 0.000 AIC: 0.473 AIC*n: 160.727 BIC: -1782.825 BIC': 30.175 ------------------------------------------------------------------------------- Full main model for HP2pbfhm for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2pbfhm ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2pbfhm subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 54.48 Prob > chi2 = 0.0000 Log likelihood = -112.37659 Pseudo R2 = 0.1951 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0764706 .0168625 4.53 0.000 .0434207 .1095204 radhlw1 | .0197767 .0060446 3.27 0.001 .0079295 .0316239 bf1 | .0102198 .0319073 0.32 0.749 -.0523175 .072757 bf20 | -.0098859 .0262026 -0.38 0.706 -.0612421 .0414703 deaw1 | -.1694401 .2100502 -0.81 0.420 -.5811308 .2422507 shhlw1 | .0028786 .0067993 0.42 0.672 -.0104478 .0162051 shjobw1 | .0109885 .0066829 1.64 0.100 -.0021097 .0240866 shrelaw1 | -.0061099 .0055939 -1.09 0.275 -.0170736 .0048539 suprtw1 | -.0018698 .0055191 -0.34 0.735 -.012687 .0089474 _cons | -7.471036 1.391934 -5.37 0.000 -10.19918 -4.742896 ------------------------------------------------------------------------------ Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 9 3 | 12 - | 38 311 | 349 -----------+--------------------------+----------- Total | 47 314 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 19.15% Specificity Pr( -|~D) 99.04% Positive predictive value Pr( D| +) 75.00% Negative predictive value Pr(~D| -) 89.11% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.96% False - rate for true D Pr( -| D) 80.85% False + rate for classified + Pr(~D| +) 25.00% False - rate for classified - Pr( D| -) 10.89% -------------------------------------------------- Correctly classified 88.64% -------------------------------------------------- Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 423.47 Prob > chi2 = 0.0039 Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -139.619 Log-Lik Full Model: -112.377 D(351): 224.753 LR(9): 54.484 Prob > LR: 0.000 McFadden's R2: 0.195 McFadden's Adj R2: 0.123 Maximum Likelihood R2: 0.140 Cragg & Uhler's R2: 0.260 McKelvey and Zavoina's R2: 0.363 Efron's R2: 0.208 Variance of y*: 5.164 Variance of error: 3.290 Count R2: 0.886 Adj Count R2: 0.128 AIC: 0.678 AIC*n: 244.753 BIC: -1842.243 BIC': -1.484 ------------------------------------------------------------------------------- Full main model for HP2sxlife for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for HP2pbfhm ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2sxlife subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 85.53 Prob > chi2 = 0.0000 Log likelihood = -128.75122 Pseudo R2 = 0.2493 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0794939 .014719 5.40 0.000 .0506451 .1083426 radhlw1 | .009177 .0050435 1.82 0.069 -.000708 .019062 bf1 | .0255037 .0309771 0.82 0.410 -.0352102 .0862176 bf20 | -.0168051 .0267683 -0.63 0.530 -.0692699 .0356598 deaw1 | .3968761 .2712323 1.46 0.143 -.1347295 .9284817 shhlw1 | .0053833 .0061644 0.87 0.383 -.0066986 .0174652 shjobw1 | .0120161 .0057139 2.10 0.035 .0008171 .0232151 shrelaw1 | -.0034493 .0049958 -0.69 0.490 -.0132409 .0063423 suprtw1 | -.001828 .0055711 -0.33 0.743 -.0127471 .0090912 _cons | -6.639149 1.272534 -5.22 0.000 -9.133269 -4.145029 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 26 16 | 42 - | 43 255 | 298 -----------+--------------------------+----------- Total | 69 271 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 37.68% Specificity Pr( -|~D) 94.10% Positive predictive value Pr( D| +) 61.90% Negative predictive value Pr(~D| -) 85.57% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.90% False - rate for true D Pr( -| D) 62.32% False + rate for classified + Pr(~D| +) 38.10% False - rate for classified - Pr( D| -) 14.43% -------------------------------------------------- Correctly classified 82.65% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 317.29 Prob > chi2 = 0.5792 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -171.514 Log-Lik Full Model: -128.751 D(330): 257.502 LR(9): 85.525 Prob > LR: 0.000 McFadden's R2: 0.249 McFadden's Adj R2: 0.191 Maximum Likelihood R2: 0.222 Cragg & Uhler's R2: 0.350 McKelvey and Zavoina's R2: 0.413 Efron's R2: 0.252 Variance of y*: 5.605 Variance of error: 3.290 Count R2: 0.826 Adj Count R2: 0.145 AIC: 0.816 AIC*n: 277.502 BIC: -1666.050 BIC': -33.065 ------------------------------------------------------------------------------- Full main model for HP2sxlife for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2sxlife ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2sxlife subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 98.58 Prob > chi2 = 0.0000 Log likelihood = -157.72791 Pseudo R2 = 0.2381 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0955211 .0143598 6.65 0.000 .0673765 .1236658 radhlw1 | .0076884 .0043462 1.77 0.077 -.00083 .0162068 bf1 | .0126394 .0277031 0.46 0.648 -.0416577 .0669365 bf20 | -.0012302 .0235195 -0.05 0.958 -.0473276 .0448672 deaw1 | .0002926 .1487155 0.00 0.998 -.2911845 .2917697 shhlw1 | -.0067303 .0056634 -1.19 0.235 -.0178304 .0043698 shjobw1 | .0163073 .0053284 3.06 0.002 .0058638 .0267508 shrelaw1 | -.0028221 .0044927 -0.63 0.530 -.0116277 .0059835 suprtw1 | .001642 .0041364 0.40 0.691 -.0064651 .0097491 _cons | -7.57151 1.189708 -6.36 0.000 -9.903294 -5.239726 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 41 20 | 61 - | 53 247 | 300 -----------+--------------------------+----------- Total | 94 267 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 43.62% Specificity Pr( -|~D) 92.51% Positive predictive value Pr( D| +) 67.21% Negative predictive value Pr(~D| -) 82.33% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 7.49% False - rate for true D Pr( -| D) 56.38% False + rate for classified + Pr(~D| +) 32.79% False - rate for classified - Pr( D| -) 17.67% -------------------------------------------------- Correctly classified 79.78% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 354.34 Prob > chi2 = 0.4105 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -207.020 Log-Lik Full Model: -157.728 D(351): 315.456 LR(9): 98.584 Prob > LR: 0.000 McFadden's R2: 0.238 McFadden's Adj R2: 0.190 Maximum Likelihood R2: 0.239 Cragg & Uhler's R2: 0.350 McKelvey and Zavoina's R2: 0.392 Efron's R2: 0.265 Variance of y*: 5.408 Variance of error: 3.290 Count R2: 0.798 Adj Count R2: 0.223 AIC: 0.929 AIC*n: 335.456 BIC: -1751.540 BIC': -45.584 ------------------------------------------------------------------------------- Full main model for HP2inthob for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for HP2sxlife ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2inthob subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 62.70 Prob > chi2 = 0.0000 Log likelihood = -87.714325 Pseudo R2 = 0.2633 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0469277 .0177122 2.65 0.008 .0122124 .0816431 radhlw1 | .0188772 .0066787 2.83 0.005 .0057871 .0319672 bf1 | -1.367019 102.0878 -0.01 0.989 -201.4554 198.7214 bf20 | 1.370719 102.0878 0.01 0.989 -198.7177 201.4591 deaw1 | -1.420629 .7247073 -1.96 0.050 -2.841029 -.000229 shhlw1 | .0101692 .007213 1.41 0.159 -.0039681 .0243065 shjobw1 | .0095172 .0070594 1.35 0.178 -.0043189 .0233533 shrelaw1 | -.0074121 .00618 -1.20 0.230 -.0195247 .0047004 suprtw1 | .0120175 .0067365 1.78 0.074 -.0011857 .0252207 _cons | -61.4171 4083.511 -0.02 0.988 -8064.952 7942.118 ------------------------------------------------------------------------------ Note: 23 failures and 0 successes completely determined. Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 8 1 | 9 - | 30 301 | 331 -----------+--------------------------+----------- Total | 38 302 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 21.05% Specificity Pr( -|~D) 99.67% Positive predictive value Pr( D| +) 88.89% Negative predictive value Pr(~D| -) 90.94% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.33% False - rate for true D Pr( -| D) 78.95% False + rate for classified + Pr(~D| +) 11.11% False - rate for classified - Pr( D| -) 9.06% -------------------------------------------------- Correctly classified 90.88% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 332.44 Prob > chi2 = 0.3467 Measures of Fit for logit of HP2inthob Log-Lik Intercept Only: -119.064 Log-Lik Full Model: -87.714 D(330): 175.429 LR(9): 62.700 Prob > LR: 0.000 McFadden's R2: 0.263 McFadden's Adj R2: 0.179 Maximum Likelihood R2: 0.168 Cragg & Uhler's R2: 0.334 McKelvey and Zavoina's R2: 0.965 Efron's R2: 0.239 Variance of y*: 94.491 Variance of error: 3.290 Count R2: 0.909 Adj Count R2: 0.184 AIC: 0.575 AIC*n: 195.429 BIC: -1748.123 BIC': -10.239 ------------------------------------------------------------------------------- Full main model for HP2inthob for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2inthob ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2inthob subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 79.74 Prob > chi2 = 0.0000 Log likelihood = -131.84212 Pseudo R2 = 0.2322 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0887071 .0158463 5.60 0.000 .0576489 .1197652 radhlw1 | .0197231 .0053656 3.68 0.000 .0092067 .0302394 bf1 | -.033217 .0371955 -0.89 0.372 -.1061189 .0396848 bf20 | .0312708 .0332387 0.94 0.347 -.0338758 .0964173 deaw1 | -.01341 .1602451 -0.08 0.933 -.3274847 .3006646 shhlw1 | .0056196 .006261 0.90 0.369 -.0066516 .0178909 shjobw1 | .0023075 .0060411 0.38 0.702 -.0095328 .0141479 shrelaw1 | -.0033791 .0049976 -0.68 0.499 -.0131743 .0064161 suprtw1 | .0018167 .0046916 0.39 0.699 -.0073787 .011012 _cons | -9.026372 1.602951 -5.63 0.000 -12.1681 -5.884647 ------------------------------------------------------------------------------ Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 26 6 | 32 - | 40 289 | 329 -----------+--------------------------+----------- Total | 66 295 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 39.39% Specificity Pr( -|~D) 97.97% Positive predictive value Pr( D| +) 81.25% Negative predictive value Pr(~D| -) 87.84% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.03% False - rate for true D Pr( -| D) 60.61% False + rate for classified + Pr(~D| +) 18.75% False - rate for classified - Pr( D| -) 12.16% -------------------------------------------------- Correctly classified 87.26% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 429.70 Prob > chi2 = 0.0020 Measures of Fit for logit of HP2inthob Log-Lik Intercept Only: -171.710 Log-Lik Full Model: -131.842 D(351): 263.684 LR(9): 79.736 Prob > LR: 0.000 McFadden's R2: 0.232 McFadden's Adj R2: 0.174 Maximum Likelihood R2: 0.198 Cragg & Uhler's R2: 0.323 McKelvey and Zavoina's R2: 0.408 Efron's R2: 0.269 Variance of y*: 5.559 Variance of error: 3.290 Count R2: 0.873 Adj Count R2: 0.303 AIC: 0.786 AIC*n: 283.684 BIC: -1803.312 BIC': -26.736 ------------------------------------------------------------------------------- Full main model for HP2vacatn for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 1 model Wave = 1 for HP2inthob ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2vacatn subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 340 LR chi2(9) = 63.50 Prob > chi2 = 0.0000 Log likelihood = -93.400079 Pseudo R2 = 0.2537 ------------------------------------------------------------------------------ HP2vacatn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0680461 .0174883 3.89 0.000 .0337697 .1023225 radhlw1 | .0124707 .0061517 2.03 0.043 .0004136 .0245278 bf1 | -.0963654 .0889765 -1.08 0.279 -.2707562 .0780254 bf20 | .0856524 .0861894 0.99 0.320 -.0832757 .2545804 deaw1 | -1.062962 .5930123 -1.79 0.073 -2.225245 .0993207 shhlw1 | .0106702 .0069711 1.53 0.126 -.0029929 .0243333 shjobw1 | .0181404 .0071032 2.55 0.011 .0042184 .0320624 shrelaw1 | -.0112469 .0061227 -1.84 0.066 -.0232473 .0007534 suprtw1 | .0083536 .0065604 1.27 0.203 -.0045047 .0212118 _cons | -10.52227 3.506457 -3.00 0.003 -17.3948 -3.649741 ------------------------------------------------------------------------------ Logistic model for HP2vacatn -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 8 4 | 12 - | 33 295 | 328 -----------+--------------------------+----------- Total | 41 299 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2vacatn != 0 -------------------------------------------------- Sensitivity Pr( +| D) 19.51% Specificity Pr( -|~D) 98.66% Positive predictive value Pr( D| +) 66.67% Negative predictive value Pr(~D| -) 89.94% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.34% False - rate for true D Pr( -| D) 80.49% False + rate for classified + Pr(~D| +) 33.33% False - rate for classified - Pr( D| -) 10.06% -------------------------------------------------- Correctly classified 89.12% -------------------------------------------------- Logistic model for HP2vacatn, goodness-of-fit test number of observations = 340 number of covariate patterns = 333 Pearson chi2(323) = 324.00 Prob > chi2 = 0.4740 Measures of Fit for logit of HP2vacatn Log-Lik Intercept Only: -125.152 Log-Lik Full Model: -93.400 D(330): 186.800 LR(9): 63.505 Prob > LR: 0.000 McFadden's R2: 0.254 McFadden's Adj R2: 0.174 Maximum Likelihood R2: 0.170 Cragg & Uhler's R2: 0.327 McKelvey and Zavoina's R2: 0.472 Efron's R2: 0.242 Variance of y*: 6.227 Variance of error: 3.290 Count R2: 0.891 Adj Count R2: 0.098 AIC: 0.608 AIC*n: 206.800 BIC: -1736.752 BIC': -11.044 ------------------------------------------------------------------------------- Full main model for HP2vacatn for wave= 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- chunk 2 H1 test:Gender= 2 model Wave = 1 for HP2vacatn ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Full Nottingham Part 2 HP2vacatn subscale models ------------------------------------------------------------------------------- Logistic regression Number of obs = 361 LR chi2(9) = 82.63 Prob > chi2 = 0.0000 Log likelihood = -125.82087 Pseudo R2 = 0.2472 ------------------------------------------------------------------------------ HP2vacatn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0884841 .0163538 5.41 0.000 .0564313 .1205368 radhlw1 | .0261977 .0057913 4.52 0.000 .014847 .0375484 bf1 | -.0344464 .0312123 -1.10 0.270 -.0956214 .0267286 bf20 | .0183328 .0265608 0.69 0.490 -.0337255 .070391 deaw1 | -.0152546 .1698279 -0.09 0.928 -.3481113 .317602 shhlw1 | -.0000247 .0063261 -0.00 0.997 -.0124237 .0123742 shjobw1 | .0075507 .0061271 1.23 0.218 -.0044583 .0195596 shrelaw1 | -.0045753 .0052069 -0.88 0.380 -.0147807 .0056301 suprtw1 | .0082219 .0046149 1.78 0.075 -.0008231 .017267 _cons | -8.575778 1.434965 -5.98 0.000 -11.38826 -5.763298 ------------------------------------------------------------------------------ Logistic model for HP2vacatn -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 23 6 | 29 - | 40 292 | 332 -----------+--------------------------+----------- Total | 63 298 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2vacatn != 0 -------------------------------------------------- Sensitivity Pr( +| D) 36.51% Specificity Pr( -|~D) 97.99% Positive predictive value Pr( D| +) 79.31% Negative predictive value Pr(~D| -) 87.95% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.01% False - rate for true D Pr( -| D) 63.49% False + rate for classified + Pr(~D| +) 20.69% False - rate for classified - Pr( D| -) 12.05% -------------------------------------------------- Correctly classified 87.26% -------------------------------------------------- Logistic model for HP2vacatn, goodness-of-fit test number of observations = 361 number of covariate patterns = 359 Pearson chi2(349) = 465.78 Prob > chi2 = 0.0000 Measures of Fit for logit of HP2vacatn Log-Lik Intercept Only: -167.134 Log-Lik Full Model: -125.821 D(351): 251.642 LR(9): 82.625 Prob > LR: 0.000 McFadden's R2: 0.247 McFadden's Adj R2: 0.187 Maximum Likelihood R2: 0.205 Cragg & Uhler's R2: 0.339 McKelvey and Zavoina's R2: 0.441 Efron's R2: 0.288 Variance of y*: 5.883 Variance of error: 3.290 Count R2: 0.873 Adj Count R2: 0.270 AIC: 0.752 AIC*n: 271.642 BIC: -1815.354 BIC': -29.626 120 . 121 . set more off 122 . *-----Chunk 2 dosew1 moderator paid employment impact------------------------ > --- 123 . title "1. H1 pt2 wv 2 male cum rad dose wrt HP2work impact " ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 1. H1 pt2 wv 2 male cum rad dose wrt HP2work impact ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:11:04 ***** ******************************************************************************* ******************************************************************************* 124 . * male models 125 . forvalues j=1/1 { 2. set more off 3. title4 "trimmed HP2work main effects models wave 1 for H1 part 2 with dose > ns" 4. title4 "Wave `j' dose HP2work relationship but avgcumdosew`j': Dose not si > gnif" 5. di _skip(2) 6. di as input "Gender =1 HP2work model" 7. logit HP2work age bf4 bf40 illw`j' movew`j' shrelaw`j' /// > avgcumdosew`j' radhlw`j' if gender==1 8. estat class 9. estat gof 10. fitstat 11. } ------------------------------------------------------------------------------- trimmed HP2work main effects models wave 1 for H1 part 2 with dose ns ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Wave 1 dose HP2work relationship but avgcumdosew1: Dose not signif ------------------------------------------------------------------------------- Gender =1 HP2work model Iteration 0: log likelihood = -172.87291 Iteration 1: log likelihood = -145.38341 Iteration 2: log likelihood = -143.32275 Iteration 3: log likelihood = -143.31043 Iteration 4: log likelihood = -143.31043 Logistic regression Number of obs = 340 LR chi2(8) = 59.12 Prob > chi2 = 0.0000 Log likelihood = -143.31043 Pseudo R2 = 0.1710 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0048446 .0141529 0.34 0.732 -.0228945 .0325838 bf4 | -.1400119 .0321951 -4.35 0.000 -.2031132 -.0769107 bf40 | .3309236 .0941317 3.52 0.000 .1464287 .5154184 illw1 | -.2284418 .3531212 -0.65 0.518 -.9205467 .4636631 movew1 | .4467728 .4500499 0.99 0.321 -.4353089 1.328854 shrelaw1 | .0013722 .0039941 0.34 0.731 -.0064562 .0092005 avgcumdosew1 | .014672 .0771675 0.19 0.849 -.1365736 .1659175 radhlw1 | -.0007224 .0043698 -0.17 0.869 -.009287 .0078422 _cons | -.8602236 .9469206 -0.91 0.364 -2.716154 .9957067 ------------------------------------------------------------------------------ Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 17 15 | 32 - | 53 255 | 308 -----------+--------------------------+----------- Total | 70 270 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 24.29% Specificity Pr( -|~D) 94.44% Positive predictive value Pr( D| +) 53.12% Negative predictive value Pr(~D| -) 82.79% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.56% False - rate for true D Pr( -| D) 75.71% False + rate for classified + Pr(~D| +) 46.88% False - rate for classified - Pr( D| -) 17.21% -------------------------------------------------- Correctly classified 80.00% -------------------------------------------------- Logistic model for HP2work, goodness-of-fit test number of observations = 340 number of covariate patterns = 335 Pearson chi2(326) = 303.46 Prob > chi2 = 0.8099 Measures of Fit for logit of HP2work Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -143.310 D(331): 286.621 LR(8): 59.125 Prob > LR: 0.000 McFadden's R2: 0.171 McFadden's Adj R2: 0.119 Maximum Likelihood R2: 0.160 Cragg & Uhler's R2: 0.250 McKelvey and Zavoina's R2: 0.254 Efron's R2: 0.166 Variance of y*: 4.413 Variance of error: 3.290 Count R2: 0.800 Adj Count R2: 0.029 AIC: 0.896 AIC*n: 304.621 BIC: -1642.760 BIC': -12.493 126 . 127 . title4 "Constructing male moderators of HP2work in wave 1" ------------------------------------------------------------------------------- Constructing male moderators of HP2work in wave 1 ------------------------------------------------------------------------------- 128 . * construction of potential moderators 129 . 130 . set more off 131 . foreach var in bf4 bf40 { 2. cap gen `var'Xd1 = `var'*avgcumdosew1 3. label var `var'Xd1 "interaction of avgcumdosew1 and `var'" 4. } 132 . 133 . 134 . 135 . 136 . des bf4 bf40 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) 137 . forvalues j=1/1 { 2. title4 "trimmed HP2work main effects models wave `j' " "male model for H1 > part 2 with dose ns" 3. title2 "Wave `j dose HP2work relationship but avgcumdosew`j': Dose not sig > nif" 4. } ------------------------------------------------------------------------------- trimmed HP2work main effects models wave 1 ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- title2: Wave `j dose HP2work relationship but avgcumdosew1: Dose not signif Date and time: 18 Jun 2012 18:11:05 Working directory: /Users/robertyaffee/ > Documents/data/research/chwk/phase3/Htests/h1tests/h1pt2 Stata data file: chwide16june2012.dta > has 2369 variables and 703 observations Wave `j dose HP2work relationship but avgcumdosew1: Dose not signif ------------------------------------------------------------------------------- 138 . 139 . 140 . set more off 141 . 142 . forvalues j=1/1 { 2. sw, pr(.1):logistic HP2work age bf4 bf40 movew`j' shrelaw` > j' /// > avgcumdosew`j' illw1 radhlw`j' bf4Xd1 bf40Xd1 if gender==1, c > oef 3. estat class 4. estat gof 5. fitstat 6. } begin with full model p = 0.8072 >= 0.1000 removing avgcumdosew1 p = 0.7761 >= 0.1000 removing shrelaw1 p = 0.7348 >= 0.1000 removing age p = 0.7348 >= 0.1000 removing radhlw1 p = 0.3215 >= 0.1000 removing movew1 p = 0.2131 >= 0.1000 removing illw1 p = 0.2607 >= 0.1000 removing bf40Xd1 p = 0.9419 >= 0.1000 removing bf4Xd1 Logistic regression Number of obs = 340 LR chi2(2) = 57.60 Prob > chi2 = 0.0000 Log likelihood = -144.07356 Pseudo R2 = 0.1666 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf40 | .3278445 .0853217 3.84 0.000 .160617 .495072 bf4 | -.1439508 .0277538 -5.19 0.000 -.1983472 -.0895544 _cons | -.5378471 .4253878 -1.26 0.206 -1.371592 .2958977 ------------------------------------------------------------------------------ Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 18 14 | 32 - | 52 256 | 308 -----------+--------------------------+----------- Total | 70 270 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 25.71% Specificity Pr( -|~D) 94.81% Positive predictive value Pr( D| +) 56.25% Negative predictive value Pr(~D| -) 83.12% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.19% False - rate for true D Pr( -| D) 74.29% False + rate for classified + Pr(~D| +) 43.75% False - rate for classified - Pr( D| -) 16.88% -------------------------------------------------- Correctly classified 80.59% -------------------------------------------------- Logistic model for HP2work, goodness-of-fit test number of observations = 340 number of covariate patterns = 91 Pearson chi2(88) = 104.61 Prob > chi2 = 0.1092 Measures of Fit for logistic of HP2work Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -144.074 D(337): 288.147 LR(2): 57.599 Prob > LR: 0.000 McFadden's R2: 0.167 McFadden's Adj R2: 0.149 Maximum Likelihood R2: 0.156 Cragg & Uhler's R2: 0.244 McKelvey and Zavoina's R2: 0.248 Efron's R2: 0.162 Variance of y*: 4.375 Variance of error: 3.290 Count R2: 0.806 Adj Count R2: 0.057 AIC: 0.865 AIC*n: 294.147 BIC: -1676.208 BIC': -45.941 143 . 144 . logit hp2work bf4 bf40 avgcumdosew1 radhlw1 havmilsq if gender==1 Iteration 0: log likelihood = -172.87291 Iteration 1: log likelihood = -146.04951 Iteration 2: log likelihood = -144.07501 Iteration 3: log likelihood = -144.06434 Iteration 4: log likelihood = -144.06434 Logistic regression Number of obs = 340 LR chi2(5) = 57.62 Prob > chi2 = 0.0000 Log likelihood = -144.06434 Pseudo R2 = 0.1666 ------------------------------------------------------------------------------ hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1443876 .0307305 -4.70 0.000 -.2046181 -.084157 bf40 | .3278933 .0856676 3.83 0.000 .1599879 .4957988 avgcumdosew1 | .0103223 .0782446 0.13 0.895 -.1430344 .1636789 radhlw1 | -.00017 .0043575 -0.04 0.969 -.0087106 .0083705 havmilsq | -5.54e-08 1.66e-06 -0.03 0.973 -3.32e-06 3.21e-06 _cons | -.5281323 .5641161 -0.94 0.349 -1.63378 .577515 ------------------------------------------------------------------------------ 145 . logit hp2work bf4 bf40 if gender==1 Iteration 0: log likelihood = -172.87291 Iteration 1: log likelihood = -146.01734 Iteration 2: log likelihood = -144.08312 Iteration 3: log likelihood = -144.07356 Iteration 4: log likelihood = -144.07356 Logistic regression Number of obs = 340 LR chi2(2) = 57.60 Prob > chi2 = 0.0000 Log likelihood = -144.07356 Pseudo R2 = 0.1666 ------------------------------------------------------------------------------ hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1439508 .0277538 -5.19 0.000 -.1983472 -.0895544 bf40 | .3278445 .0853217 3.84 0.000 .160617 .495072 _cons | -.5378471 .4253878 -1.26 0.206 -1.371592 .2958977 ------------------------------------------------------------------------------ 146 . estat class Logistic model for hp2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 18 14 | 32 - | 52 256 | 308 -----------+--------------------------+----------- Total | 70 270 | 340 Classified + if predicted Pr(D) >= .5 True D defined as hp2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 25.71% Specificity Pr( -|~D) 94.81% Positive predictive value Pr( D| +) 56.25% Negative predictive value Pr(~D| -) 83.12% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.19% False - rate for true D Pr( -| D) 74.29% False + rate for classified + Pr(~D| +) 43.75% False - rate for classified - Pr( D| -) 16.88% -------------------------------------------------- Correctly classified 80.59% -------------------------------------------------- 147 . estat gof Logistic model for hp2work, goodness-of-fit test number of observations = 340 number of covariate patterns = 91 Pearson chi2(88) = 104.61 Prob > chi2 = 0.1092 148 . fitstat Measures of Fit for logit of hp2work Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -144.074 D(337): 288.147 LR(2): 57.599 Prob > LR: 0.000 McFadden's R2: 0.167 McFadden's Adj R2: 0.149 Maximum Likelihood R2: 0.156 Cragg & Uhler's R2: 0.244 McKelvey and Zavoina's R2: 0.248 Efron's R2: 0.162 Variance of y*: 4.375 Variance of error: 3.290 Count R2: 0.806 Adj Count R2: 0.057 AIC: 0.865 AIC*n: 294.147 BIC: -1676.208 BIC': -45.941 149 . 150 . * capturing significant vars from last analysis 151 . local cn1: colnames(e(b)) 152 . di "`cn1'" bf4 bf40 _cons 153 . local leng1 = length( "`cn1'") 154 . di `leng1' 14 155 . local leng1b `leng1'-6 156 . di `leng1b' 8 157 . local nuvlist = substr("`cn1'",1,`leng1b') 158 . di "`nuvlist'" bf4 bf40 159 . local rhsvars = "`nuvlist'" 160 . local nuvlist= "`nuvlist'" 161 . local nuvlist= substr("`cn1'",1,`leng1b') 162 . di "`nuvlist'" bf4 bf40 163 . sw, pr(.1):logit hp2hmcare `nuvlist' if gender==1 begin with full model p < 0.1000 for all terms in model Logistic regression Number of obs = 340 LR chi2(2) = 81.32 Prob > chi2 = 0.0000 Log likelihood = -132.21538 Pseudo R2 = 0.2352 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.2248287 .0309809 -7.26 0.000 -.2855501 -.1641072 bf40 | .1658314 .0893229 1.86 0.063 -.0092382 .340901 _cons | .7314062 .4552436 1.61 0.108 -.1608548 1.623667 ------------------------------------------------------------------------------ 164 . di "`rhsvars'" bf4 bf40 165 . matrix define c=e(b) 166 . local cn2: colnames(c) 167 . di "`cn2'" bf4 bf40 _cons 168 . local leng2 = length("`cn2'") 169 . local leng2b = `leng2'-6 170 . local rhsvars = substr("`cn2'",1,`leng2b') 171 . logit hp2work `rhsvars' if gender==1 Iteration 0: log likelihood = -172.87291 Iteration 1: log likelihood = -146.01734 Iteration 2: log likelihood = -144.08312 Iteration 3: log likelihood = -144.07356 Iteration 4: log likelihood = -144.07356 Logistic regression Number of obs = 340 LR chi2(2) = 57.60 Prob > chi2 = 0.0000 Log likelihood = -144.07356 Pseudo R2 = 0.1666 ------------------------------------------------------------------------------ hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1439508 .0277538 -5.19 0.000 -.1983472 -.0895544 bf40 | .3278445 .0853217 3.84 0.000 .160617 .495072 _cons | -.5378471 .4253878 -1.26 0.206 -1.371592 .2958977 ------------------------------------------------------------------------------ 172 . 173 . 174 . 175 . di "`rhsvars'" bf4 bf40 176 . local varlist2 =substr("`rhsvars'",1,9) 177 . di "`varlist2'" bf4 bf40 178 . 179 . * constructing potential moderators 180 . foreach var in bf4 bf40 { 2. cap gen `var'Xd1 = `var'* avgcumdosew1 3. } 181 . 182 . *x no signif male moderators for paid employment 183 . set more off 184 . sw, pr(.1): logistic hp2work `rhsvars' bf4Xd1 bf40Xd1 if gender==1, coef begin with full model p = 0.2607 >= 0.1000 removing bf40Xd1 p = 0.9419 >= 0.1000 removing bf4Xd1 Logistic regression Number of obs = 340 LR chi2(2) = 57.60 Prob > chi2 = 0.0000 Log likelihood = -144.07356 Pseudo R2 = 0.1666 ------------------------------------------------------------------------------ hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1439508 .0277538 -5.19 0.000 -.1983472 -.0895544 bf40 | .3278445 .0853217 3.84 0.000 .160617 .495072 _cons | -.5378471 .4253878 -1.26 0.206 -1.371592 .2958977 ------------------------------------------------------------------------------ 185 . fitstat Measures of Fit for logistic of hp2work Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -144.074 D(337): 288.147 LR(2): 57.599 Prob > LR: 0.000 McFadden's R2: 0.167 McFadden's Adj R2: 0.149 Maximum Likelihood R2: 0.156 Cragg & Uhler's R2: 0.244 McKelvey and Zavoina's R2: 0.248 Efron's R2: 0.162 Variance of y*: 4.375 Variance of error: 3.290 Count R2: 0.806 Adj Count R2: 0.057 AIC: 0.865 AIC*n: 294.147 BIC: -1676.208 BIC': -45.941 186 . 187 . scalar wkModMw1 = "none" 188 . di _skip(2) 189 . title4 "testing the female moderator model Hp2work H1 Pt 2 wave 1" ------------------------------------------------------------------------------- testing the female moderator model Hp2work H1 Pt 2 wave 1 ------------------------------------------------------------------------------- 190 . * Testing female moderator model xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx > xxx 191 . 192 . title4 "testing general female moderator model for hp2work" ------------------------------------------------------------------------------- testing general female moderator model for hp2work ------------------------------------------------------------------------------- 193 . local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 194 . di _skip(4) 195 . 196 . 197 . forvalues j=1/1 { 2. set more off 3. di _skip(4) 4. di as input "For females hp2work on wave 1 with dose ns" 5. des age occ1w`j'-occ8w`j' inc1w`j'-inc4w`j' avgcumdosew`j' `w1bf' 6. sw, pr(.1): logistic HP2work age havmilsq /// > avgcumdosew1 illw`j' shjobw`j' suprtw`j' radhlw`j' if gender==2, coef 7. estat gof 8. estat class 9. fitstat 10. } For females hp2work on wave 1 with dose ns storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age occ1w1 double %15.0g LABJ profess executive administration in 1986 occ2w1 double %15.0g LABJ technical sales admin support in 1986 occ3w1 double %15.0g LABJ service occup protective services in 1986 occ4w1 double %15.0g LABJ precision prod mechan craft construction in 1986 occ5w1 double %15.0g LABJ factory laborer machinist transp cleaner in 1986 occ6w1 double %15.0g LABJ farming agricul forestry fishing trapping logging in 1986 occ7w1 double %15.0g LABJ homemaking or caregiving in 1986 occ8w1 double %15.0g LABJ student in 1986 inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf10 float %9.0g bf10= max(0, sufamw1 - 20) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf20 float %9.0g bf20 = max(0, kzchorn - 2.53946E-006) bf22 float %9.0g bf22 = max(0, icdxcnt - 1.01635E-007) * bf7m bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) begin with full model p = 0.9219 >= 0.1000 removing suprtw1 p = 0.8867 >= 0.1000 removing shjobw1 p = 0.6189 >= 0.1000 removing havmilsq p = 0.1241 >= 0.1000 removing avgcumdosew1 Logistic regression Number of obs = 361 LR chi2(3) = 46.25 Prob > chi2 = 0.0000 Log likelihood = -182.84519 Pseudo R2 = 0.1123 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0626548 .0119743 5.23 0.000 .0391855 .086124 radhlw1 | .0071024 .0037129 1.91 0.056 -.0001748 .0143795 illw1 | .4077261 .2382216 1.71 0.087 -.0591797 .8746319 _cons | -4.859841 .6646145 -7.31 0.000 -6.162461 -3.55722 ------------------------------------------------------------------------------ Logistic model for HP2work, goodness-of-fit test number of observations = 361 number of covariate patterns = 253 Pearson chi2(249) = 271.11 Prob > chi2 = 0.1605 Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 24 6 | 30 - | 69 262 | 331 -----------+--------------------------+----------- Total | 93 268 | 361 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 25.81% Specificity Pr( -|~D) 97.76% Positive predictive value Pr( D| +) 80.00% Negative predictive value Pr(~D| -) 79.15% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.24% False - rate for true D Pr( -| D) 74.19% False + rate for classified + Pr(~D| +) 20.00% False - rate for classified - Pr( D| -) 20.85% -------------------------------------------------- Correctly classified 79.22% -------------------------------------------------- Measures of Fit for logistic of HP2work Log-Lik Intercept Only: -205.969 Log-Lik Full Model: -182.845 D(357): 365.690 LR(3): 46.247 Prob > LR: 0.000 McFadden's R2: 0.112 McFadden's Adj R2: 0.093 Maximum Likelihood R2: 0.120 Cragg & Uhler's R2: 0.177 McKelvey and Zavoina's R2: 0.192 Efron's R2: 0.141 Variance of y*: 4.072 Variance of error: 3.290 Count R2: 0.792 Adj Count R2: 0.194 AIC: 1.035 AIC*n: 373.690 BIC: -1736.639 BIC': -28.580 198 . 199 . cap gen ageXd1 = age*avgcumdosew1 200 . 201 . * capturing significant vars 202 . local cn3: colnames(e(b)) 203 . di "`cn3'" age radhlw1 illw1 _cons 204 . local leng3 = length( "`cn3'") 205 . di `leng3' 23 206 . local leng3b `leng3'-6 207 . di `leng3b' 17 208 . local nuvlist3 = substr("`cn3'",1,`leng3b') 209 . di "`nuvlist3'" age radhlw1 illw1 210 . local rhsvars3 = "`nuvlist3'" 211 . local rhsvars4= substr("`cn3'",1,`leng3b') 212 . di "`rhsvars4'" age radhlw1 illw1 213 . 214 . * moderators for hp2work female and male are saved as scalars: 215 . scalar wkModFw1="none" 216 . scalar wkModMw1="none" 217 . cap gen ageXd1 = age*avgcumdosew1 218 . 219 . *x no significant female moderator for paid employment 220 . forvalues j=1/1 { 2. di as input "For females hp2probsoc on wave 1 with dose ns" 3. des age occ1w`j'-occ8w`j' inc1w`j'-inc4w`j' avgcumdosew`j' 4. set more off 5. sw, pr(.1): logistic HP2work bf4 bf14 age havmil /// > avgcumdosew1 ageXd1 illw`j' accdw`j' suprtw`j' if gender==2, coef 6. estat gof 7. estat class 8. fitstat 9. } For females hp2probsoc on wave 1 with dose ns storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age occ1w1 double %15.0g LABJ profess executive administration in 1986 occ2w1 double %15.0g LABJ technical sales admin support in 1986 occ3w1 double %15.0g LABJ service occup protective services in 1986 occ4w1 double %15.0g LABJ precision prod mechan craft construction in 1986 occ5w1 double %15.0g LABJ factory laborer machinist transp cleaner in 1986 occ6w1 double %15.0g LABJ farming agricul forestry fishing trapping logging in 1986 occ7w1 double %15.0g LABJ homemaking or caregiving in 1986 occ8w1 double %15.0g LABJ student in 1986 inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 begin with full model p = 0.9043 >= 0.1000 removing suprtw1 p = 0.5398 >= 0.1000 removing accdw1 p = 0.4992 >= 0.1000 removing havmil p = 0.3576 >= 0.1000 removing illw1 p = 0.2119 >= 0.1000 removing ageXd1 p = 0.1822 >= 0.1000 removing avgcumdosew1 p = 0.2326 >= 0.1000 removing bf14 Logistic regression Number of obs = 362 LR chi2(2) = 59.51 Prob > chi2 = 0.0000 Log likelihood = -176.51284 Pseudo R2 = 0.1442 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1117282 .0265443 -4.21 0.000 -.163754 -.0597024 age | .0511864 .01239 4.13 0.000 .0269024 .0754704 _cons | -2.684594 .7848001 -3.42 0.001 -4.222774 -1.146414 ------------------------------------------------------------------------------ Logistic model for HP2work, goodness-of-fit test number of observations = 362 number of covariate patterns = 274 Pearson chi2(271) = 303.76 Prob > chi2 = 0.0834 Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 31 18 | 49 - | 62 251 | 313 -----------+--------------------------+----------- Total | 93 269 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 33.33% Specificity Pr( -|~D) 93.31% Positive predictive value Pr( D| +) 63.27% Negative predictive value Pr(~D| -) 80.19% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 6.69% False - rate for true D Pr( -| D) 66.67% False + rate for classified + Pr(~D| +) 36.73% False - rate for classified - Pr( D| -) 19.81% -------------------------------------------------- Correctly classified 77.90% -------------------------------------------------- Measures of Fit for logistic of HP2work Log-Lik Intercept Only: -206.266 Log-Lik Full Model: -176.513 D(359): 353.026 LR(2): 59.506 Prob > LR: 0.000 McFadden's R2: 0.144 McFadden's Adj R2: 0.130 Maximum Likelihood R2: 0.152 Cragg & Uhler's R2: 0.223 McKelvey and Zavoina's R2: 0.234 Efron's R2: 0.175 Variance of y*: 4.295 Variance of error: 3.290 Count R2: 0.779 Adj Count R2: 0.140 AIC: 0.992 AIC*n: 359.026 BIC: -1762.075 BIC': -47.723 221 . 222 . scalar SigDoseFw1 = "no" 223 . scalar MainEffwkFw1 = "bf4 age" 224 . ****** Moderator analysis for Dose=>paid employment wave one 225 . title4 "testing potential moderators for women and hp2work in wave 1" ------------------------------------------------------------------------------- testing potential moderators for women and hp2work in wave 1 ------------------------------------------------------------------------------- 226 . set more off 227 . forvalues j=1/1 { 2. sw, pr(.1):logistic HP2work age occ1w1-occ8w1 bf8 illw`j' > shjobw`j' havmilsq /// > avgcumdosew`j' ageXd1 if gender==2, coef 3. estat class 4. estat gof 5. fitstat 6. } begin with full model p = 0.9407 >= 0.1000 removing occ4w1 p = 0.7880 >= 0.1000 removing shjobw1 p = 0.7709 >= 0.1000 removing bf8 p = 0.7819 >= 0.1000 removing occ7w1 p = 0.7309 >= 0.1000 removing occ3w1 p = 0.6557 >= 0.1000 removing havmilsq p = 0.3777 >= 0.1000 removing occ6w1 p = 0.4099 >= 0.1000 removing occ2w1 p = 0.2835 >= 0.1000 removing occ1w1 p = 0.1994 >= 0.1000 removing occ5w1 p = 0.1230 >= 0.1000 removing ageXd1 p = 0.1760 >= 0.1000 removing illw1 p = 0.1245 >= 0.1000 removing occ8w1 p = 0.1089 >= 0.1000 removing avgcumdosew1 Logistic regression Number of obs = 363 LR chi2(1) = 41.64 Prob > chi2 = 0.0000 Log likelihood = -185.74123 Pseudo R2 = 0.1008 ------------------------------------------------------------------------------ HP2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0695394 .011616 5.99 0.000 .0467724 .0923064 _cons | -4.709665 .6434135 -7.32 0.000 -5.970733 -3.448598 ------------------------------------------------------------------------------ Logistic model for HP2work -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 20 14 | 34 - | 73 256 | 329 -----------+--------------------------+----------- Total | 93 270 | 363 Classified + if predicted Pr(D) >= .5 True D defined as HP2work != 0 -------------------------------------------------- Sensitivity Pr( +| D) 21.51% Specificity Pr( -|~D) 94.81% Positive predictive value Pr( D| +) 58.82% Negative predictive value Pr(~D| -) 77.81% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.19% False - rate for true D Pr( -| D) 78.49% False + rate for classified + Pr(~D| +) 41.18% False - rate for classified - Pr( D| -) 22.19% -------------------------------------------------- Correctly classified 76.03% -------------------------------------------------- Logistic model for HP2work, goodness-of-fit test number of observations = 363 number of covariate patterns = 49 Pearson chi2(47) = 58.17 Prob > chi2 = 0.1274 Measures of Fit for logistic of HP2work Log-Lik Intercept Only: -206.563 Log-Lik Full Model: -185.741 D(361): 371.482 LR(1): 41.643 Prob > LR: 0.000 McFadden's R2: 0.101 McFadden's Adj R2: 0.091 Maximum Likelihood R2: 0.108 Cragg & Uhler's R2: 0.159 McKelvey and Zavoina's R2: 0.172 Efron's R2: 0.125 Variance of y*: 3.971 Variance of error: 3.290 Count R2: 0.760 Adj Count R2: 0.065 AIC: 1.034 AIC*n: 375.482 BIC: -1756.397 BIC': -35.748 228 . 229 . * capturing significant vars 230 . local cn5: colnames(e(b)) 231 . di "`cn5'" age _cons 232 . local leng5 = length( "`cn5'") 233 . di `leng5' 9 234 . local leng5b `leng5'-6 235 . di `leng5b' 3 236 . local nuvlist5 = substr("`cn5'",1,`leng5b') 237 . di "`nuvlist5'" age 238 . local rhsvars5 = "`nuvlist2'" 239 . local nuvlist6= "`nuvlist2'" 240 . local nuvlist6= substr("`cn5'",1,`leng5b') 241 . di "`nuvlist6'" age 242 . 243 . foreach varx in `nuvlist6' { 2. cap gen `varx'X`vary' = `varx'*avgcumdosew1 3. } 244 . 245 . cap gen illw1Xd1 = illw1*avgcumdosew1 246 . 247 . logit hp2hmcare age avgcumdosew1 illw1 ageXd1 illw1Xd1 if gender==2 > // correct regress on previous Iteration 0: log likelihood = -233.72859 Iteration 1: log likelihood = -189.61734 Iteration 2: log likelihood = -188.76377 Iteration 3: log likelihood = -188.76057 Iteration 4: log likelihood = -188.76057 Logistic regression Number of obs = 363 LR chi2(5) = 89.94 Prob > chi2 = 0.0000 Log likelihood = -188.76057 Pseudo R2 = 0.1924 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0970647 .0160129 6.06 0.000 .0656799 .1284495 avgcumdosew1 | .884954 2.296761 0.39 0.700 -3.616615 5.386523 illw1 | .9418439 .3176563 2.96 0.003 .319249 1.564439 ageXd1 | -.0212632 .0412659 -0.52 0.606 -.1021429 .0596166 illw1Xd1 | -.2742628 .427414 -0.64 0.521 -1.111979 .5634532 _cons | -5.742044 .8635113 -6.65 0.000 -7.434495 -4.049593 ------------------------------------------------------------------------------ 248 . 249 . 250 . scalar SigDoseWkFw1 = "no" 251 . scalar SigDoseWkMw1 = "no" 252 . des bf4 bf40 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) 253 . scalar MainEffwkMw1 = "bf4 bf40" 254 . scalar MainEffwkFw1 = "age " 255 . scalar WKModMw1 = "none" 256 . scalar WkModFw1 = "ageXd1" 257 . 258 . 259 . * male sign main effects in main effects model: 2- bf4, bf40 260 . * male and female main effects model avgcumdosew1 were not signif. 261 . * male hp2wk w1 mediators: bf4 and bf40 262 . * female signif main effects in main effects model 263 . 264 . title4 "H1 pt 2 wave 1 Mediation of paid employment testing for males" ------------------------------------------------------------------------------- H1 pt 2 wave 1 Mediation of paid employment testing for males ------------------------------------------------------------------------------- 265 . 266 . * male hp2wk w1 mediators: testing b4 and b40 267 . 268 . cap gen ageXillw1 = age*illw1 269 . correlate bf4 age if gender==1 (obs=340) | bf4 age -------------+------------------ bf4 | 1.0000 age | -0.4041 1.0000 270 . 271 . 272 . glm age avgcumdosew1 if gender==1, fam(gauss) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 273 . glm hp2work age if gender==1, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 332.6127 Iteration 2: deviance = 332.2035 Iteration 3: deviance = 332.2034 Iteration 4: deviance = 332.2034 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 332.2034085 (1/df) Deviance = .9828503 Pearson = 339.5831812 (1/df) Pearson = 1.004684 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1637.98 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0233394 .0063554 3.67 0.000 .0108831 .0357956 _cons | -2.001321 .3351159 -5.97 0.000 -2.658136 -1.344506 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 274 . 275 . 276 . des bf4 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) 277 . glm bf4 avgcumdosew1 if gender==1, fam(gauss) link(identity) Iteration 0: log likelihood = -1026.9659 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 24.75428 Deviance = 8366.946191 (1/df) Deviance = 24.75428 Pearson = 8366.946191 (1/df) Pearson = 24.75428 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.05274 Log likelihood = -1026.965868 BIC = 6396.763 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.1031788 .161925 -0.64 0.524 -.4205461 .2141884 _cons | 12.54134 .2786337 45.01 0.000 11.99523 13.08746 ------------------------------------------------------------------------------ 278 . glm hp2work bf4 if gender==1, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 302.5414 Iteration 2: deviance = 301.6138 Iteration 3: deviance = 301.6078 Iteration 4: deviance = 301.6078 Iteration 5: deviance = 301.6078 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 301.6078179 (1/df) Deviance = .8923308 Pearson = 318.2302284 (1/df) Pearson = .9415096 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1668.576 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1012401 .0146623 -6.90 0.000 -.1299776 -.0725026 _cons | .3470813 .1825598 1.90 0.057 -.0107293 .704892 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 279 . 280 . 281 . 282 . 283 . des bf40 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) 284 . glm bf40 avgcumdosew1 if gender==1, fam(gauss) link(identity) Iteration 0: log likelihood = -659.60303 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 2.852057 Deviance = 963.9952523 (1/df) Deviance = 2.852057 Pearson = 963.9952523 (1/df) Pearson = 2.852057 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 3.891783 Log likelihood = -659.6030291 BIC = -1006.188 ------------------------------------------------------------------------------ | OIM bf40 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0453656 .0549627 0.83 0.409 -.0623593 .1530904 _cons | 2.124647 .0945774 22.46 0.000 1.939278 2.310015 ------------------------------------------------------------------------------ 285 . glm hp2work bf40 if gender==1, fam(binomial) link(probit) /// > irls scale(dev) Iteration 1: deviance = 316.1036 Iteration 2: deviance = 315.5033 Iteration 3: deviance = 315.5021 Iteration 4: deviance = 315.5021 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 315.5020956 (1/df) Deviance = .9334382 Pearson = 333.9910612 (1/df) Pearson = .9881392 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1654.682 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf40 | .2493852 .044782 5.57 0.000 .161614 .3371563 _cons | -1.421911 .1366145 -10.41 0.000 -1.689671 -1.154152 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 286 . title4 "bf40 could be mediator for males h2wk in wave 1" ------------------------------------------------------------------------------- bf40 could be mediator for males h2wk in wave 1 ------------------------------------------------------------------------------- 287 . 288 . 289 . 290 . glm illw1 avgcumdosew1 if gender==1, fam(gaussian) link(identity) Iteration 0: log likelihood = -151.48261 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = .1435742 Deviance = 48.52808533 (1/df) Deviance = .1435742 Pearson = 48.52808533 (1/df) Pearson = .1435742 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = .9028389 Log likelihood = -151.4826069 BIC = -1921.656 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0087277 .0123318 0.71 0.479 -.0154423 .0328976 _cons | .0962541 .0212201 4.54 0.000 .0546635 .1378446 ------------------------------------------------------------------------------ 291 . glm hp2work illw1 if gender==1, fam(binomial) irls /// > scale(dev) link(probit) Iteration 1: deviance = 345.5327 Iteration 2: deviance = 345.2889 Iteration 3: deviance = 345.2888 Iteration 4: deviance = 345.2888 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 345.2887981 (1/df) Deviance = 1.021564 Pearson = 340.1227725 (1/df) Pearson = 1.00628 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1624.895 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .1303424 .1949057 0.67 0.504 -.2516658 .5123505 _cons | -.8347401 .080778 -10.33 0.000 -.993062 -.6764181 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 292 . 293 . scalar WkMedMw1 = "bf40" 294 . 295 . 296 . 297 . title4 "Testing possible female paid employment mediators for paid employment > in wave 1" ------------------------------------------------------------------------------- Testing possible female paid employment mediators for paid employment in wave 1 ------------------------------------------------------------------------------- 298 . 299 . title4 "Test of age as possible female mediator in wave 1" ------------------------------------------------------------------------------- Test of age as possible female mediator in wave 1 ------------------------------------------------------------------------------- 300 . glm age avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 301 . glm hp2work age if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 372.7391 Iteration 2: deviance = 372.3711 Iteration 3: deviance = 372.3711 Iteration 4: deviance = 372.3711 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 372.3710546 (1/df) Deviance = 1.031499 Pearson = 375.1783727 (1/df) Pearson = 1.039275 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1755.508 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0398718 .0066503 6.00 0.000 .0268375 .052906 _cons | -2.722762 .3594297 -7.58 0.000 -3.427231 -2.018292 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 302 . title4 "age can be a wave 1 mediator for women with hp2work " ------------------------------------------------------------------------------- age can be a wave 1 mediator for women with hp2work ------------------------------------------------------------------------------- 303 . 304 . 305 . title4 "Test of b40 as female mediator of paid employment in wave 1" ------------------------------------------------------------------------------- Test of b40 as female mediator of paid employment in wave 1 ------------------------------------------------------------------------------- 306 . glm bf40 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -818.49948 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 5.351018 Deviance = 1931.717442 (1/df) Deviance = 5.351018 Pearson = 1931.717442 (1/df) Pearson = 5.351018 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 4.520658 Log likelihood = -818.4994757 BIC = -196.162 ------------------------------------------------------------------------------ | OIM bf40 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .4524905 .2213302 2.04 0.041 .0186913 .8862898 _cons | 3.013471 .1423225 21.17 0.000 2.734524 3.292417 ------------------------------------------------------------------------------ 307 . glm hp2work bf40 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 406.4865 Iteration 2: deviance = 406.0549 Iteration 3: deviance = 406.0547 Iteration 4: deviance = 406.0547 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 406.0547444 (1/df) Deviance = 1.124805 Pearson = 360.1809662 (1/df) Pearson = .9977312 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1721.825 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf40 | .080565 .0316866 2.54 0.011 .0184603 .1426696 _cons | -.921639 .1303745 -7.07 0.000 -1.177168 -.6661097 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 308 . title4 "b40 could be a wv 1 mediator for women" ------------------------------------------------------------------------------- b40 could be a wv 1 mediator for women ------------------------------------------------------------------------------- 309 . 310 . 311 . title4 "bf4 Test of female mediation of paid employment in wave 1" ------------------------------------------------------------------------------- bf4 Test of female mediation of paid employment in wave 1 ------------------------------------------------------------------------------- 312 . glm bf4 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 313 . glm hp2work bf4 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 370.9717 Iteration 2: deviance = 370.7179 Iteration 3: deviance = 370.7176 Iteration 4: deviance = 370.7176 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 370.7176349 (1/df) Deviance = 1.026919 Pearson = 356.8979308 (1/df) Pearson = .9886369 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1757.162 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.090794 .0144985 -6.26 0.000 -.1192106 -.0623775 _cons | .2257867 .1560342 1.45 0.148 -.0800347 .531608 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 314 . title4 "bf4 alone is a mediator for women" ------------------------------------------------------------------------------- bf4 alone is a mediator for women ------------------------------------------------------------------------------- 315 . 316 . 317 . title "bf4 Test of female mediation of paid employment in wave 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** bf4 Test of female mediation of paid employment in wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:12:37 ***** ******************************************************************************* ******************************************************************************* 318 . glm bf4 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 319 . glm hp2work bf4 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 370.9717 Iteration 2: deviance = 370.7179 Iteration 3: deviance = 370.7176 Iteration 4: deviance = 370.7176 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 370.7176349 (1/df) Deviance = 1.026919 Pearson = 356.8979308 (1/df) Pearson = .9886369 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1757.162 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.090794 .0144985 -6.26 0.000 -.1192106 -.0623775 _cons | .2257867 .1560342 1.45 0.148 -.0800347 .531608 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 320 . title4 "bf4 could be a mediator for women" ------------------------------------------------------------------------------- bf4 could be a mediator for women ------------------------------------------------------------------------------- 321 . 322 . 323 . 324 . title "Test of female mediation of illw1 and paid employment in wave 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Test of female mediation of illw1 and paid employment in wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:12:40 ***** ******************************************************************************* ******************************************************************************* 325 . glm illw1 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -259.70777 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = .2462383 Deviance = 88.89203958 (1/df) Deviance = .2462383 Pearson = 88.89203958 (1/df) Pearson = .2462383 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 1.441916 Log likelihood = -259.7077741 BIC = -2038.987 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0275426 .1585711 _cons | .1570822 .0305304 5.15 0.000 .0972436 .2169207 ------------------------------------------------------------------------------ 326 . glm hp2work illw1 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 408.8941 Iteration 2: deviance = 408.3909 Iteration 3: deviance = 408.3907 Iteration 4: deviance = 408.3907 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 408.3907209 (1/df) Deviance = 1.131276 Pearson = 362.2934789 (1/df) Pearson = 1.003583 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1719.489 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .294665 .1449437 2.03 0.042 .0105805 .5787495 _cons | -.7143685 .0815701 -8.76 0.000 -.8742431 -.554494 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 327 . title4 "illw1 alone is a mediator for women" ------------------------------------------------------------------------------- illw1 alone is a mediator for women ------------------------------------------------------------------------------- 328 . 329 . 330 . 331 . 332 . title "Test of female mediation of paid employment in wave 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Test of female mediation of paid employment in wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:12:42 ***** ******************************************************************************* ******************************************************************************* 333 . mvreg bf4 age bf40 = avgcumdosew1 if gender==2 Equation Obs Parms RMSE "R-sq" F P ---------------------------------------------------------------------- bf4 363 2 5.149837 0.0253 9.376729 0.0024 age 363 2 11.6814 0.0338 12.64139 0.0004 bf40 363 2 2.313227 0.0114 4.179627 0.0416 ------------------------------------------------------------------------------ | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.477832 -.5398375 _cons | 10.99384 .3168463 34.70 0.000 10.37075 11.61694 -------------+---------------------------------------------------------------- age | avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.775898 6.171859 _cons | 48.88157 .7187038 68.01 0.000 47.4682 50.29494 -------------+---------------------------------------------------------------- bf40 | avgcumdosew1 | .4524905 .2213302 2.04 0.042 .017232 .887749 _cons | 3.013471 .1423225 21.17 0.000 2.733585 3.293356 ------------------------------------------------------------------------------ 334 . glm hp2work bf4 age bf40 if gender==2, fam(binomial) link(probit) irls /// > scale(dev) Iteration 1: deviance = 354.2404 Iteration 2: deviance = 353.2021 Iteration 3: deviance = 353.1978 Iteration 4: deviance = 353.1978 Iteration 5: deviance = 353.1978 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 359 (IRLS EIM) Scale parameter = 1 Deviance = 353.1977844 (1/df) Deviance = .9838378 Pearson = 373.3187421 (1/df) Pearson = 1.039885 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1762.893 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.0639211 .0157071 -4.07 0.000 -.0947065 -.0331358 age | .0282268 .007046 4.01 0.000 .0144169 .0420367 bf40 | .0293818 .0331211 0.89 0.375 -.0355344 .0942981 _cons | -1.597741 .4643893 -3.44 0.001 -2.507928 -.6875549 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 335 . title4 "age & b4 together could be wave 1 mediators for women" ------------------------------------------------------------------------------- age & b4 together could be wave 1 mediators for women ------------------------------------------------------------------------------- 336 . 337 . 338 . 339 . 340 . title "Test of female mediation of paid employment in wave 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Test of female mediation of paid employment in wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:12:45 ***** ******************************************************************************* ******************************************************************************* 341 . mvreg bf4 age bf40 illw1 = avgcumdosew1 if gender==2 Equation Obs Parms RMSE "R-sq" F P ---------------------------------------------------------------------- bf4 363 2 5.149837 0.0253 9.376729 0.0024 age 363 2 11.6814 0.0338 12.64139 0.0004 bf40 363 2 2.313227 0.0114 4.179627 0.0416 illw1 363 2 .4962241 0.0052 1.904017 0.1685 ------------------------------------------------------------------------------ | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.477832 -.5398375 _cons | 10.99384 .3168463 34.70 0.000 10.37075 11.61694 -------------+---------------------------------------------------------------- age | avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.775898 6.171859 _cons | 48.88157 .7187038 68.01 0.000 47.4682 50.29494 -------------+---------------------------------------------------------------- bf40 | avgcumdosew1 | .4524905 .2213302 2.04 0.042 .017232 .887749 _cons | 3.013471 .1423225 21.17 0.000 2.733585 3.293356 -------------+---------------------------------------------------------------- illw1 | avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0278557 .1588841 _cons | .1570822 .0305304 5.15 0.000 .0970423 .217122 ------------------------------------------------------------------------------ 342 . glm hp2work bf4 age bf40 illw1 if gender==2, fam(binomial) /// > link(probit) irls scale(dev) Iteration 1: deviance = 354.001 Iteration 2: deviance = 352.9698 Iteration 3: deviance = 352.9654 Iteration 4: deviance = 352.9654 Iteration 5: deviance = 352.9654 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 358 (IRLS EIM) Scale parameter = 1 Deviance = 352.965421 (1/df) Deviance = .9859369 Pearson = 374.1304084 (1/df) Pearson = 1.045057 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1757.231 ------------------------------------------------------------------------------ | EIM hp2work | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.0628029 .0158842 -3.95 0.000 -.0939353 -.0316706 age | .0280305 .0070628 3.97 0.000 .0141877 .0418734 bf40 | .0276234 .0334236 0.83 0.409 -.0378856 .0931324 illw1 | .0692009 .1432947 0.48 0.629 -.2116515 .3500534 _cons | -1.60672 .4655573 -3.45 0.001 -2.519195 -.6942446 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 343 . qui: { When all are together only age and b4 are a wave 1 mediators for women 344 . 345 . 346 . scalar wkMedMw1 = "bf40 " 347 . 348 . scalar SigDoseWkMw1 = "no" 349 . scalar MainEffwkMw1 = "age" 350 . scalar MainEffwkFw1 = "age" 351 . scalar wkMedFw1 = "age b4" 352 . * male hp2wk w1 mediators: testing b4 and b40 353 . 354 . title4 "*---1.- Summary matrix construction Paid employment partition: first > two rows" ------------------------------------------------------------------------------- *---1.- Summary matrix construction Paid employment partition: first two rows ------------------------------------------------------------------------------- 355 . 356 . matrix define wkMw1 = J(1,8, 0) 357 . matrix define wkFw1 = J(1,8, 0) 358 . matrix colnames wkMw1= hypnum ptnum wave gender medsig numMAsig numModsig n > umMed 359 . matrix colnames wkFw1= hypnum ptnum wave gender medsig numMAsig numModsig n > umMed 360 . matrix rownames wkMw1 = workMw1 361 . matrix rownames wkFw1 = workFw1 362 . matrix define wkMw1= (1, 2, 1, 1, 0 ,2, 0 , 1 ) 363 . matrix define wkFw1= (1, 2, 1, 2, 0, 1, 1 , 2 ) 364 . 365 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1) 366 . matrix colnames H1pt2w1 = hypnum ptnum wave medsig numMAsig numModsig n > umMed 367 . matrix rownames H1pt2w1 = wkMw1 wkFw1 368 . matlist H1pt2w1 | hypnum ptnum wave medsig numMAsig numModsig > numMed numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 369 . 370 . set more off 371 . scalar list MainEffwkFw1 = age MainEffwkMw1 = age MainEffVactnMw1 = age radhlw1 VactnMedFw1 = age illw1 radhlw1 VactnMedMw1 = age VacatnModFw1 = none MainEffVactnFw1 = age radhlw1 bf7m SigDoseVactnFw1 = no VactnModMw1 = none inthobMedFw1 = age bf4 illw1 bf4m inthobMedMw1 = age inthobMw1 = age InthbModFw1 = none MainEffInthbFw1 = age radhlw1 bf4 SigdoseInthbFw1 = no InthbModMw1 = none MainEffInthbMw1 = age radhlw1 shfamw1 SigDoseInthbMw1 = no MainEffMw1 = radhlw1 bf4 bf40 sxlifeMedMw1 = radhlw1 sxlifeMedFw1 = age illw1 radhlw1 bf4 bf4m MainEffsxlifeFw1 = age bf4 bf4m SigDoseSxlifeFw1 = no SigDosesxlifeMw1 = no MainEffsxlifeMw1 = age bf4 bf40 MainEffhmcrFw1 = age MainEffhmcrMw1 = age hmcrMedFw1 = age bf4 hmcrMedMw1 = radhlw1 SigDosePrbfmhmMw1 = no vactnModMw1 = none SigDoseVactnMw1 = no SxLifeModFw1 = no sxlifeModFw1 = none sxlifeModMw1 = none MaineffhmcrMw1 = age bf4 bf40 SigDoseMEhmcrW1 = no PrbfmhmMedFw1 = age bf4 PrbfmhmMedMw1 = age MainEffPrbfmhmFw1 = age radhlw1 bf4 MainEffPrbfmhmMw1 = age bf4 PrbfmhmModFw1 = none PrbfmhmModMw1 = none SigDosePrbfmhmFw1 = no SigDosePrbfhmMw1 = no MainEffPrbfhmMw1 = age bf4 MainEffVactnMw2 = age radhlw2 sxlifeMedMw2 = age illw2 SigDoseSxlifeFw2 = no MainEffsxlifeFw2 = age radhlw2 bf4 bf4m MainEffPrbsocMw2 = age radhlw2 shjobw2 MainEffhmcrFw2 = age hmcrMedFw2 = age bf4 MainEffwkFw2 = age MainEffwkMw2 = age MainEffPrbsocMw1 = age bf4m SigdoseMw1 = no ProbsocMedFw1 = age bf4 ProbsocMedMw1 = radhlw1 ProbsocModFw1 = none SigDoseProbsocMw1 = no hmcrmedMw1 = radhlw1 hmcrmedFw1 = age b4 b40 SigdosehmcrFw1 = no MainEffProbSocFw1 = age radhlw1 avgcumdosew1 shrelaw1 bf4 SigDoseProbsocFw1 = yes PrbsocModMw1 = shjobw1Xd1 shrelaw1Xd1 WkhmcrMw1 = age b4 WkModFw1 = ageXd1 hmcareMedFw1 = age illw1 hmcareMedMw1 = age SigDosehmcrFw1 = no wkMedMw1 = bf40 hmcrModFw1 = none SigDoseHmcrFw1 = yes WkMedMw1 = bf40 hmcrModMw1 = none SigDosehmcrMw1 = no wkMedFw1 = age b4 WKModMw1 = none SigDoseWkMw1 = no SigDoseWkFw1 = no SigDoseFw1 = no wkModFw1 = none wkModMw1 = none VactnMedMw2 = age inthobMedMw2 = age inthobMw2 = age PrbfmhmModMw2 = none MainEffProbSocFw2 = age radhlw2 avgcumdosew2 bf4 hmcrModMw2 = none MainEffhmcrMw2 = age wkMedFw2 = age b4 wkMedMw2 = age bf4 MainEffsxlifeMw2 = age bf4 bf40 shjobw2 shrelaw2 radhlw2 MainEffPrbfmhmMw2 = bf4 bf6 bf7 ProbsocMedFw2 = age bf4 radhlw2 hmcareMedFw2 = age bf4 WkhmcrMw2 = age b4 MainEffhmcrw2 = age hmcrModFw2 = none SigDoseHmcrFw2 = yes NumhmcrModMw2 = none SigDosehmcrMw2 = no SigdosehmcrFw2 = yes hmcrMedMw2 = age ageXillw2 SigDosehmcrFw2 = no MainEffhmcareMw2 = age WkMedMw2 = age ageXillw2 wkMedFw3 = radhlw3 age ageXillw3 bf40 bf4m bf1 VactnMedFw2 = age illw2 radhlw2 VacatnModFw2 = none MainEffVactnFw2 = age radhlw2 bf7m SigDoseVactnFw2 = no VactnModMw2 = none vactnModMw2 = none SigDoseVactnMw2 = no inthobMedFw2 = age bf4 illw2 bf4m InthbModFw2 = none MainEffInthbFw2 = age radhlw2 bf4 SigdoseInthbFw2 = no InthbModMw2 = none MainEffInthbMw2 = age radhlw2 shfamw2 SigDoseInthbMw2 = no MainEffMw2 = radhlw2 bf4 bf40 SigdoseMEinthob = no sxlifeMedFw2 = age illw2 radhlw2 bf4 bf4m SxLifeModFw2 = no sxlifeModFw2 = none sxlifeModMw2 = none SigDosesxlifeMw2 = no PrbfmhmMedFw2 = age bf4 PrbfmhmMedMw2 = age PrbfmhmModFw2 = none MainEffPrbfmhmFw2 = age bf4 bf40 SigDosePrbfmhmFw2 = no PrbfmhmModw2 = none SigDosePrbfmhmMw2 = no SigDosePrbfhmMw2 = no MainEffPrbfhmMw2 = bf4 bf6 bf7 ProbsocMedMw2 = age ProbsocModFw2 = none SigDoseProbsocFw2 = yes ProbSocModMw2 = none SigDoseProbsocMw2 = no PrbsocModMw2 = none SigdoseMw2 = none hmcareMedMw2 = age hmcareModFw2 = none MainEffhmcarew2 = age SigdoseHmcareFw2 = no hmcareModMw2 = none SigDoseHmcareMw2 = no NameMedMw2 = age ageXillw2 NumModMw2 = none SigDosehmcareMw2 = no SigDoseWKMw2 = no WkMedFw2 = age bf4 WkModFw2 = none WKModMw2 = none SigDoseWkMw2 = no SigDoseWkFw2 = no SigDoseFw2 = no wkModFw2 = none wkModMw2 = none VactnMedFw3 = age illw3 radhlw3 VactnMedMw3 = age illw3 VacatnModFw3 = none MainEffVactnFw3 = age radhlw3 deaw3 SigDoseVactnFw3 = no vactnModMw3 = none MainEffVactnMw3 = age bf7m radhlw3 SigDoseVactnMw3 = no sxLifeMedFw3 = age bf4 bf4m sxLifeMedMw3 = age illw3 InthbModFw3 = none MainEffInthbFw3 = age radhlw3 bf4 SigdoseInthbFw3 = no InthbMw3 = none MainEffInthbMw3 = age radhlw3 shfamw3 SigDoseInthbMw3 = no sxlifeMedFw3 = age illw3 radhlw3 bf4 bf4m sxlifeMedMw3 = age illw3 sxlifeModFw3 = none MainEffsxlifeFw3 = age radhlw3 bf4 bf4m shrelaw3 shfamw3 SigDoseSxlifeFw3 = no sxlifeModMw3 = none SigDosesxlifeMw3 = no MainEffsxlifeMw3 = age bf4 illw3 radhlw3 PrbfmhmMedFw3 = age bf4 PrbfmhmMedMw3 = age PrbfmhmModFw3 = none MainEffPrbfmhmFw3 = age bf4 bf40 SigDosePrbfmhmFw3 = no PrbfmhmModw3 = none SigDosePrbfmhmMw3 = no SigDosePrbfhmMw3 = no MainEffPrbfhmMw3 = bf1 bf4 dvcew3 bf7m ProbsocMedFw3 = age radhlw3 ProbsocMedMw3 = age ProbsocModFw3 = none MainEffProbSocFw3 = age radhlw3 illw3 Shrelaw3 avgcumodsew3 SigDoseProbsocFw3 = yes ProbSocModMw3 = none SigDoseProbsocMw3 = no MainEffPrbsocMw3 = age radhlw3 shjobw3 hmcareMedFw3 = age illw3 hmcareMedMw3 = age illw3 hmcareModFw3 = none SigdoseHmcareFw3 = no hmcareModMw3 = none MainEffhmcareMw3 = none SigDoseHmcareMw3 = no wkMedMw3 = bf8 age illw3 ageXillw3 wkModFw3 = none wkModMw3 = none MainEffwkFw3 = age MainEffwkMw3 = workM: age bf8 illw3 shjobw3 SigDoseWKMw3 = no SigDoseWkFw3 = no 372 . * moderator construction 373 . * none ussed because basic dose work relationship washes out 374 . 375 . title "2. Hyp 1 pt 2 wave 1 male dose Hp2hmcare impact explored" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 2. Hyp 1 pt 2 wave 1 male dose Hp2hmcare impact explored ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:12:48 ***** ******************************************************************************* ******************************************************************************* 376 . 377 . * ------- testing male and female moderators-wave 1 for hmcare-------------- > --- 378 . 379 . cap gen hp2hmcare=HP2hmcare 380 . 381 . forvalues j=1/1 { 2. set more off 3. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 4. di as input "For females hp2hmcare on wave 1 with dose ns" 5. des age occ1w`j'-occ8w`j' inc1w`j'-inc4w`j' avgcumdosew`j' bf8 // > / > marrw`j'1-marrw`j'6 `w1bf' radhlw`j' 6. sw, pr(.1): logistic hp2hmcare marrw`j'3-marrw`j'6 age havmilsq / > // > avgcumdosew1 bf8 illw`j' shjobw`j' suprtw`j' if gender==1, co > ef 7. estat gof 8. estat class 9. fitstat 10. } For females hp2hmcare on wave 1 with dose ns storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age occ1w1 double %15.0g LABJ profess executive administration in 1986 occ2w1 double %15.0g LABJ technical sales admin support in 1986 occ3w1 double %15.0g LABJ service occup protective services in 1986 occ4w1 double %15.0g LABJ precision prod mechan craft construction in 1986 occ5w1 double %15.0g LABJ factory laborer machinist transp cleaner in 1986 occ6w1 double %15.0g LABJ farming agricul forestry fishing trapping logging in 1986 occ7w1 double %15.0g LABJ homemaking or caregiving in 1986 occ8w1 double %15.0g LABJ student in 1986 inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 bf8 float %9.0g bf8 = max(0, radtlw3 - 40) * bf5m marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf10 float %9.0g bf10= max(0, sufamw1 - 20) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf20 float %9.0g bf20 = max(0, kzchorn - 2.53946E-006) bf22 float %9.0g bf22 = max(0, icdxcnt - 1.01635E-007) * bf7m bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) radhlw1 double %8.0g how much believed personal health is affected by radiation in 1986 note: marrw14 dropped because of collinearity note: marrw16 dropped because of collinearity begin with full model p = 0.8518 >= 0.1000 removing marrw13 p = 0.6179 >= 0.1000 removing avgcumdosew1 p = 0.5000 >= 0.1000 removing havmilsq p = 0.5276 >= 0.1000 removing bf8 p = 0.5433 >= 0.1000 removing illw1 p = 0.4921 >= 0.1000 removing marrw15 p = 0.1693 >= 0.1000 removing suprtw1 Logistic regression Number of obs = 340 LR chi2(2) = 41.30 Prob > chi2 = 0.0000 Log likelihood = -152.22537 Pseudo R2 = 0.1194 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0525652 .0122122 4.30 0.000 .0286297 .0765007 shjobw1 | .0156714 .0041469 3.78 0.000 .0075436 .0237992 _cons | -4.957356 .7134246 -6.95 0.000 -6.355643 -3.55907 ------------------------------------------------------------------------------ Logistic model for hp2hmcare, goodness-of-fit test number of observations = 340 number of covariate patterns = 215 Pearson chi2(212) = 221.53 Prob > chi2 = 0.3127 Logistic model for hp2hmcare -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 14 4 | 18 - | 56 266 | 322 -----------+--------------------------+----------- Total | 70 270 | 340 Classified + if predicted Pr(D) >= .5 True D defined as hp2hmcare != 0 -------------------------------------------------- Sensitivity Pr( +| D) 20.00% Specificity Pr( -|~D) 98.52% Positive predictive value Pr( D| +) 77.78% Negative predictive value Pr(~D| -) 82.61% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.48% False - rate for true D Pr( -| D) 80.00% False + rate for classified + Pr(~D| +) 22.22% False - rate for classified - Pr( D| -) 17.39% -------------------------------------------------- Correctly classified 82.35% -------------------------------------------------- Measures of Fit for logistic of hp2hmcare Log-Lik Intercept Only: -172.873 Log-Lik Full Model: -152.225 D(337): 304.451 LR(2): 41.295 Prob > LR: 0.000 McFadden's R2: 0.119 McFadden's Adj R2: 0.102 Maximum Likelihood R2: 0.114 Cragg & Uhler's R2: 0.179 McKelvey and Zavoina's R2: 0.217 Efron's R2: 0.130 Variance of y*: 4.204 Variance of error: 3.290 Count R2: 0.824 Adj Count R2: 0.143 AIC: 0.913 AIC*n: 310.451 BIC: -1659.904 BIC': -29.637 382 . 383 . cap gen shjobw1Xd1 = shjobw1*avgcumdosew1 384 . 385 . logit hp2hmcare age avgcumdosew1 shjobw1 ageXd1 shjobw1Xd1 if gender==1 Iteration 0: log likelihood = -172.87291 Iteration 1: log likelihood = -152.72837 Iteration 2: log likelihood = -151.52069 Iteration 3: log likelihood = -151.51061 Iteration 4: log likelihood = -151.51059 Iteration 5: log likelihood = -151.51059 Logistic regression Number of obs = 340 LR chi2(5) = 42.72 Prob > chi2 = 0.0000 Log likelihood = -151.51059 Pseudo R2 = 0.1236 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0459744 .0152127 3.02 0.003 .0161581 .0757907 avgcumdosew1 | -2.020359 2.276313 -0.89 0.375 -6.481851 2.441132 shjobw1 | .0138632 .0047443 2.92 0.003 .0045646 .0231618 ageXd1 | .0271336 .0358792 0.76 0.449 -.0431882 .0974555 shjobw1Xd1 | .0056837 .007561 0.75 0.452 -.0091355 .020503 _cons | -4.439055 .908991 -4.88 0.000 -6.220645 -2.657466 ------------------------------------------------------------------------------ 386 . 387 . scalar SigDosehmcrMw1 = "no" 388 . scalar MainEffhmcrMw1 = "age shjobw1" 389 . scalar hmcrModMw1 = "none" 390 . 391 . 392 . 393 . title4 "Trimmed female Hp2hmcare moderator model" ------------------------------------------------------------------------------- Trimmed female Hp2hmcare moderator model ------------------------------------------------------------------------------- 394 . 395 . cap gen ageXd1= age*avgcumdosew1 396 . 397 . forvalues j=1/1 { 2. set more off 3. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 4. di as input "For females HP2hmcare on wave 1 with dose ns" 5. des age avgcumdosew`j' /// > marrw`j'1-marrw`j'6 `w1bf' radhlw`j' 6. logistic hp2hmcare marrw`j'3-marrw`j'6 age havmilsq /// > avgcumdosew1 illw`j' ageXd1 if gender==2, coef 7. estat gof 8. estat class 9. fitstat 10. } For females HP2hmcare on wave 1 with dose ns storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf10 float %9.0g bf10= max(0, sufamw1 - 20) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf20 float %9.0g bf20 = max(0, kzchorn - 2.53946E-006) bf22 float %9.0g bf22 = max(0, icdxcnt - 1.01635E-007) * bf7m bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) radhlw1 double %8.0g how much believed personal health is affected by radiation in 1986 note: marrw14 != 0 predicts success perfectly marrw14 dropped and 3 obs not used Logistic regression Number of obs = 360 LR chi2(8) = 90.30 Prob > chi2 = 0.0000 Log likelihood = -185.35656 Pseudo R2 = 0.1959 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- marrw13 | -.3849014 .3723136 -1.03 0.301 -1.114623 .3448198 marrw14 | 0 (omitted) marrw15 | -.9153623 1.029513 -0.89 0.374 -2.93317 1.102445 marrw16 | -.9459556 1.17984 -0.80 0.423 -3.258399 1.366488 age | .1113504 .0191706 5.81 0.000 .0737767 .1489241 havmilsq | -2.13e-06 2.61e-06 -0.82 0.413 -7.24e-06 2.97e-06 avgcumdosew1 | 1.596621 2.0241 0.79 0.430 -2.370543 5.563785 illw1 | .8779977 .2787263 3.15 0.002 .3317043 1.424291 ageXd1 | -.0367027 .0350736 -1.05 0.295 -.1054456 .0320402 _cons | -6.146477 .9034853 -6.80 0.000 -7.917275 -4.375678 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. Logistic model for hp2hmcare, goodness-of-fit test number of observations = 360 number of covariate patterns = 338 Pearson chi2(329) = 329.08 Prob > chi2 = 0.4884 Logistic model for hp2hmcare -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 61 28 | 89 - | 61 210 | 271 -----------+--------------------------+----------- Total | 122 238 | 360 Classified + if predicted Pr(D) >= .5 True D defined as hp2hmcare != 0 -------------------------------------------------- Sensitivity Pr( +| D) 50.00% Specificity Pr( -|~D) 88.24% Positive predictive value Pr( D| +) 68.54% Negative predictive value Pr(~D| -) 77.49% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 11.76% False - rate for true D Pr( -| D) 50.00% False + rate for classified + Pr(~D| +) 31.46% False - rate for classified - Pr( D| -) 22.51% -------------------------------------------------- Correctly classified 75.28% -------------------------------------------------- Measures of Fit for logistic of hp2hmcare Log-Lik Intercept Only: -230.506 Log-Lik Full Model: -185.357 D(350): 370.713 LR(8): 90.300 Prob > LR: 0.000 McFadden's R2: 0.196 McFadden's Adj R2: 0.152 Maximum Likelihood R2: 0.222 Cragg & Uhler's R2: 0.307 McKelvey and Zavoina's R2: 0.468 Efron's R2: 0.243 Variance of y*: 6.185 Variance of error: 3.290 Count R2: 0.753 Adj Count R2: 0.270 AIC: 1.085 AIC*n: 390.713 BIC: -1689.423 BIC': -43.211 398 . 399 . title4 "Trimmed model of hp2work for women in wave 1" ------------------------------------------------------------------------------- Trimmed model of hp2work for women in wave 1 ------------------------------------------------------------------------------- 400 . forvalues j=1/1 { 2. set more off 3. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 4. di as input "For females HP2hmcare on wave 1 with dose ns" 5. des age avgcumdosew`j' /// > marrw`j'1-marrw`j'6 `w1bf' radhlw`j' 6. logistic hp2hmcare age avgcumdosew1 illw`j' if gender==2, coef 7. estat gof 8. estat class 9. fitstat 10. } For females HP2hmcare on wave 1 with dose ns storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf10 float %9.0g bf10= max(0, sufamw1 - 20) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf20 float %9.0g bf20 = max(0, kzchorn - 2.53946E-006) bf22 float %9.0g bf22 = max(0, icdxcnt - 1.01635E-007) * bf7m bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) radhlw1 double %8.0g how much believed personal health is affected by radiation in 1986 Logistic regression Number of obs = 363 LR chi2(3) = 88.59 Prob > chi2 = 0.0000 Log likelihood = -189.43609 Pseudo R2 = 0.1895 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0919881 .0122684 7.50 0.000 .0679425 .1160337 avgcumdosew1 | -.5411658 .2831823 -1.91 0.056 -1.096193 .0138613 illw1 | .8054954 .2675545 3.01 0.003 .2810981 1.329893 _cons | -5.38915 .6535341 -8.25 0.000 -6.670053 -4.108247 ------------------------------------------------------------------------------ Logistic model for hp2hmcare, goodness-of-fit test number of observations = 363 number of covariate patterns = 313 Pearson chi2(309) = 323.98 Prob > chi2 = 0.2677 Logistic model for hp2hmcare -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 64 28 | 92 - | 61 210 | 271 -----------+--------------------------+----------- Total | 125 238 | 363 Classified + if predicted Pr(D) >= .5 True D defined as hp2hmcare != 0 -------------------------------------------------- Sensitivity Pr( +| D) 51.20% Specificity Pr( -|~D) 88.24% Positive predictive value Pr( D| +) 69.57% Negative predictive value Pr(~D| -) 77.49% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 11.76% False - rate for true D Pr( -| D) 48.80% False + rate for classified + Pr(~D| +) 30.43% False - rate for classified - Pr( D| -) 22.51% -------------------------------------------------- Correctly classified 75.48% -------------------------------------------------- Measures of Fit for logistic of hp2hmcare Log-Lik Intercept Only: -233.729 Log-Lik Full Model: -189.436 D(359): 378.872 LR(3): 88.585 Prob > LR: 0.000 McFadden's R2: 0.190 McFadden's Adj R2: 0.172 Maximum Likelihood R2: 0.217 Cragg & Uhler's R2: 0.299 McKelvey and Zavoina's R2: 0.304 Efron's R2: 0.241 Variance of y*: 4.725 Variance of error: 3.290 Count R2: 0.755 Adj Count R2: 0.288 AIC: 1.066 AIC*n: 386.872 BIC: -1737.218 BIC': -70.902 401 . 402 . qui { ------------------------------------------------------------------------------- Caveat: We drop bf8 because it is dependent on a variable in a future wave. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- If it depended on a past wave variable we might have kept it. ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- But when it is dependent on something that has not happened yet, ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- it is too latent to consider in a wave specific analysis. ------------------------------------------------------------------------------- 403 . des bf8 bf5m bf4m storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf8 float %9.0g bf8 = max(0, radtlw3 - 40) * bf5m bf5m float %9.0g bf5m = max(0, ecprw3 - 75) * bf4m bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) 404 . scalar SigDosehmcrFw1 = "no" 405 . scalar MainEffhmcrFw1 = "illw1 age" 406 . 407 . 408 . title4 "home care wave 1 male mediator analysis" ------------------------------------------------------------------------------- home care wave 1 male mediator analysis ------------------------------------------------------------------------------- 409 . 410 . title4 "age is a possible male mediator of home care in wave 1" ------------------------------------------------------------------------------- age is a possible male mediator of home care in wave 1 ------------------------------------------------------------------------------- 411 . glm age avgcumdosew1 if gender==1, fam(gaussian) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 412 . glm hp2hmcare age if gender==1, fam(binomial) irls scale(dev) link(probit) Iteration 1: deviance = 320.8295 Iteration 2: deviance = 320.0337 Iteration 3: deviance = 320.0328 Iteration 4: deviance = 320.0328 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 320.0327771 (1/df) Deviance = .9468425 Pearson = 341.699231 (1/df) Pearson = 1.010944 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1650.151 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .032545 .0064212 5.07 0.000 .0199596 .0451304 _cons | -2.483665 .3435565 -7.23 0.000 -3.157023 -1.810306 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 413 . 414 . scalar hmcrMedMw1 = "age ageXillw1" 415 . 416 . di as input "age is a possible male mediator for home care in wave 1" age is a possible male mediator for home care in wave 1 417 . 418 . 419 . di as input "age and illw1 as main effects together suppress illw1" age and illw1 as main effects together suppress illw1 420 . glm illw1 age avgcumdosew1 if gender==1, family(gaussian) link(identity) Iteration 0: log likelihood = -149.77052 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 337 Scale parameter = .1425573 Deviance = 48.04180784 (1/df) Deviance = .1425573 Pearson = 48.04180784 (1/df) Pearson = .1425573 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = .8986501 Log likelihood = -149.7705231 BIC = -1916.313 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0031119 .0016849 1.85 0.065 -.0001905 .0064143 avgcumdosew1 | .0066365 .0123401 0.54 0.591 -.0175497 .0308227 _cons | -.0558998 .0850529 -0.66 0.511 -.2226005 .1108009 ------------------------------------------------------------------------------ 421 . glm hp2hmcare illw1 age if gender==1, family(binomial) irls scale(dev) link( > probit) Iteration 1: deviance = 320.7839 Iteration 2: deviance = 319.9774 Iteration 3: deviance = 319.9764 Iteration 4: deviance = 319.9764 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 337 (IRLS EIM) Scale parameter = 1 Deviance = 319.9764494 (1/df) Deviance = .949485 Pearson = 341.9238822 (1/df) Pearson = 1.014611 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1644.378 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .0470226 .1922185 0.24 0.807 -.3297186 .4237639 age | .032421 .0064525 5.02 0.000 .0197743 .0450676 _cons | -2.482523 .34428 -7.21 0.000 -3.1573 -1.807747 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 422 . 423 . 424 . di as input "Interaction of age and wave 1 illness for males is mediator-mode > rator on hmcare" Interaction of age and wave 1 illness for males is mediator-moderator on hmcare 425 . 426 . cap gen ageXillw1 = age*illw1 427 . glm illw1 avgcumdosew1 if gender==1, family(gaussian) link(identity) Iteration 0: log likelihood = -151.48261 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = .1435742 Deviance = 48.52808533 (1/df) Deviance = .1435742 Pearson = 48.52808533 (1/df) Pearson = .1435742 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = .9028389 Log likelihood = -151.4826069 BIC = -1921.656 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0087277 .0123318 0.71 0.479 -.0154423 .0328976 _cons | .0962541 .0212201 4.54 0.000 .0546635 .1378446 ------------------------------------------------------------------------------ 428 . glm hp2hmcare illw1 avgcumdosew1 ageXillw1 if gender==1, family(binomial) /// > irls scale(dev) link(probit) Iteration 1: deviance = 344.1427 Iteration 2: deviance = 343.92 Iteration 3: deviance = 343.9199 Iteration 4: deviance = 343.9199 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 336 (IRLS EIM) Scale parameter = 1 Deviance = 343.9198791 (1/df) Deviance = 1.023571 Pearson = 339.000371 (1/df) Pearson = 1.00893 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1614.606 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | -1.213319 1.209268 -1.00 0.316 -3.583441 1.156803 avgcumdosew1 | -.0042013 .0484889 -0.09 0.931 -.0992378 .0908352 ageXillw1 | .0242393 .0213349 1.14 0.256 -.0175764 .0660551 _cons | -.8321418 .0831923 -10.00 0.000 -.9951958 -.6690878 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 429 . 430 . 431 . 432 . * Saving mediators as scalars for Dose=work relationship 433 . **** female mediators of dose-paid employment: radhlw1, age, bf40, bf4m, bf4, > bf1 434 . 435 . scalar hmcrMedMw1 = "age" 436 . scalar hmcrMedFw1 = "age avgcumdosew1 bf4 bf4m " 437 . 438 . cap gen hp2vacatn = HP2vacatn 439 . 440 . 441 . title "2. H1 pt2 wave 1 Dose female moderator Homecare impact " ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 2. H1 pt2 wave 1 Dose female moderator Homecare impact ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:13:14 ***** ******************************************************************************* ******************************************************************************* 442 . 443 . 444 . * review of general model for men and women 445 . 446 . forvalues j=1/1 { 2. set more off 3. 447 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 448 . foreach var in HP2hmcare { 5. 449 . local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 6. title "chunk 3 H1 test pt 2 :Gender= `k' model Wave = `j' for `e(dep > var)' " 7. di _skip(4) 8. 450 . 451 . xi: logistic `var' age i.educ occ1w`j'-occ8w`j' /// > marrw`j'1- marrw`j'3 marrw`j'5-marrw`j'6 inc1w`j'-inc4w`j' // > / > radhlw`j' havmil avgcumdosew`j' `w`j'bf' /// > deaw`j' dvcew`j' sepaw`j' accdw`j' movew`j' radhlw1 /// > illw`j' shfamw`j' shhlw`j' shjobw`j' shrelaw`j' suprtw`j' suc > hrw`j' /// > havmilsq if gender==2, coef difficult iterate(50) 9. estat class 10. estat gof 11. fitstat 12. 452 . } 13. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** chunk 3 H1 test pt 2 :Gender= model Wave = 1 for hp2hmcare ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:13:14 ***** ******************************************************************************* ******************************************************************************* i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: _Ieduc_8 omitted because of collinearity note: radhlw1 omitted because of collinearity Logistic regression Number of obs = 358 LR chi2(51) = 163.36 Prob > chi2 = 0.0000 Log likelihood = -149.9205 Pseudo R2 = 0.3527 ------------------------------------------------------------------------------ HP2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1322255 .0241652 5.47 0.000 .0848625 .1795885 _Ieduc_2 | -10.0193 754.0186 -0.01 0.989 -1487.869 1467.83 _Ieduc_3 | -10.49152 754.0185 -0.01 0.989 -1488.341 1467.358 _Ieduc_4 | -8.71423 754.0188 -0.01 0.991 -1486.564 1469.135 _Ieduc_5 | -10.4068 754.0186 -0.01 0.989 -1488.256 1467.443 _Ieduc_6 | -10.68819 754.0185 -0.01 0.989 -1488.537 1467.161 _Ieduc_7 | -10.94334 754.0203 -0.01 0.988 -1488.796 1466.909 _Ieduc_8 | 0 (omitted) occ1w1 | -2.443994 1.415397 -1.73 0.084 -5.218121 .3301329 occ2w1 | -2.519671 1.458446 -1.73 0.084 -5.378173 .3388315 occ3w1 | -2.372771 1.446929 -1.64 0.101 -5.2087 .4631575 occ4w1 | -3.256872 1.570864 -2.07 0.038 -6.335709 -.1780355 occ5w1 | -3.906156 1.725894 -2.26 0.024 -7.288845 -.5234666 occ6w1 | -5.192057 1.956306 -2.65 0.008 -9.026346 -1.357767 occ7w1 | -.9645317 1.49363 -0.65 0.518 -3.891993 1.962929 occ8w1 | -1.541951 1.448029 -1.06 0.287 -4.380037 1.296134 marrw11 | -.4640303 1.925158 -0.24 0.810 -4.237272 3.309211 marrw12 | -1.711412 2.266816 -0.75 0.450 -6.15429 2.731467 marrw13 | -1.34516 1.914926 -0.70 0.482 -5.098345 2.408026 marrw15 | -2.648889 2.280264 -1.16 0.245 -7.118125 1.820346 marrw16 | -3.140349 2.467603 -1.27 0.203 -7.976762 1.696063 inc1w1 | 1.997721 1.384475 1.44 0.149 -.7157992 4.711242 inc2w1 | 2.374508 1.354203 1.75 0.080 -.2796802 5.028697 inc3w1 | 2.719588 1.365246 1.99 0.046 .0437547 5.395421 inc4w1 | 2.169417 1.522601 1.42 0.154 -.8148263 5.15366 radhlw1 | .000015 .0054568 0.00 0.998 -.0106801 .0107101 havmil | -.003359 .0026195 -1.28 0.200 -.008493 .0017751 avgcumdosew1 | -.9556895 .3881039 -2.46 0.014 -1.716359 -.1950199 bf1 | -.0087777 .0246959 -0.36 0.722 -.0571808 .0396254 bf4 | -.5899724 .228594 -2.58 0.010 -1.038008 -.1419365 bf9 | -.0436396 .0223962 -1.95 0.051 -.0875353 .0002562 bf10 | .0090231 .0130434 0.69 0.489 -.0165414 .0345877 bf11 | .0532598 .0430698 1.24 0.216 -.0311554 .1376749 bf4m | .4051708 .2117859 1.91 0.056 -.0099218 .8202635 bf15m | .0002493 .0003321 0.75 0.453 -.0004015 .0009002 bf20 | -.0059985 .0202682 -0.30 0.767 -.0457235 .0337264 bf22 | -.0000321 .0000692 -0.46 0.643 -.0001678 .0001036 bf30 | .0002449 .0002901 0.84 0.398 -.0003236 .0008135 bf40 | .0632557 .0911174 0.69 0.488 -.115331 .2418425 deaw1 | .158171 .2197751 0.72 0.472 -.2725802 .5889222 dvcew1 | .1254408 2.091225 0.06 0.952 -3.973284 4.224166 sepaw1 | -.6497514 2.414062 -0.27 0.788 -5.381225 4.081722 accdw1 | .2831415 .8314301 0.34 0.733 -1.346432 1.912715 movew1 | .2074149 .4471555 0.46 0.643 -.6689937 1.083823 radhlw1 | 0 (omitted) illw1 | 1.106215 .3492383 3.17 0.002 .4217202 1.790709 shfamw1 | -.0002582 .0064446 -0.04 0.968 -.0128893 .0123729 shhlw1 | -.006468 .0095563 -0.68 0.499 -.025198 .012262 shjobw1 | .0002575 .007117 0.04 0.971 -.0136915 .0142066 shrelaw1 | -.0148903 .0065942 -2.26 0.024 -.0278146 -.001966 suprtw1 | -.0041909 .0088489 -0.47 0.636 -.0215344 .0131526 suchrw1 | .0319242 .0187143 1.71 0.088 -.0047551 .0686035 havmilsq | 5.48e-07 1.88e-06 0.29 0.771 -3.14e-06 4.23e-06 _cons | 3.542841 754.0254 0.00 0.996 -1474.32 1481.405 ------------------------------------------------------------------------------ Logistic model for HP2hmcare -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 85 27 | 112 - | 40 206 | 246 -----------+--------------------------+----------- Total | 125 233 | 358 Classified + if predicted Pr(D) >= .5 True D defined as HP2hmcare != 0 -------------------------------------------------- Sensitivity Pr( +| D) 68.00% Specificity Pr( -|~D) 88.41% Positive predictive value Pr( D| +) 75.89% Negative predictive value Pr(~D| -) 83.74% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 11.59% False - rate for true D Pr( -| D) 32.00% False + rate for classified + Pr(~D| +) 24.11% False - rate for classified - Pr( D| -) 16.26% -------------------------------------------------- Correctly classified 81.28% -------------------------------------------------- Logistic model for HP2hmcare, goodness-of-fit test number of observations = 358 number of covariate patterns = 358 Pearson chi2(306) = 314.35 Prob > chi2 = 0.3590 Measures of Fit for logistic of HP2hmcare Log-Lik Intercept Only: -231.600 Log-Lik Full Model: -149.920 D(304): 299.841 LR(51): 163.358 Prob > LR: 0.000 McFadden's R2: 0.353 McFadden's Adj R2: 0.120 Maximum Likelihood R2: 0.366 Cragg & Uhler's R2: 0.505 McKelvey and Zavoina's R2: 0.599 Efron's R2: 0.401 Variance of y*: 8.211 Variance of error: 3.290 Count R2: 0.813 Adj Count R2: 0.464 AIC: 1.139 AIC*n: 407.841 BIC: -1487.841 BIC': 136.549 453 . 454 . title4 "Partly trimmed Female Main effects model for dose=> homecare wave 1" ------------------------------------------------------------------------------- Partly trimmed Female Main effects model for dose=> homecare wave 1 ------------------------------------------------------------------------------- 455 . logit hp2hmcare age radhlw1 /// > avgcumdosew1 bf4 if gender==2 Iteration 0: log likelihood = -233.30573 Iteration 1: log likelihood = -184.19159 Iteration 2: log likelihood = -183.06271 Iteration 3: log likelihood = -183.05768 Iteration 4: log likelihood = -183.05768 Logistic regression Number of obs = 362 LR chi2(4) = 100.50 Prob > chi2 = 0.0000 Log likelihood = -183.05768 Pseudo R2 = 0.2154 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0770328 .0130012 5.93 0.000 .0515508 .1025148 radhlw1 | -.0011136 .0035739 -0.31 0.755 -.0081184 .0058911 avgcumdosew1 | -.5567889 .2722015 -2.05 0.041 -1.090294 -.0232837 bf4 | -.1227731 .0262991 -4.67 0.000 -.1743185 -.0712277 _cons | -3.173011 .7808961 -4.06 0.000 -4.703539 -1.642482 ------------------------------------------------------------------------------ 456 . 457 . title4 "Partly trimmed Female Main effects model for dose=> homecare wave 1" > ------------------------------------------------------------------------------- Partly trimmed Female Main effects model for dose=> homecare wave 1 ------------------------------------------------------------------------------- 458 . logit hp2hmcare age avgcumdosew1 bf4 if gender==2 Iteration 0: log likelihood = -233.72859 Iteration 1: log likelihood = -184.31143 Iteration 2: log likelihood = -183.15831 Iteration 3: log likelihood = -183.15306 Iteration 4: log likelihood = -183.15306 Logistic regression Number of obs = 363 LR chi2(3) = 101.15 Prob > chi2 = 0.0000 Log likelihood = -183.15306 Pseudo R2 = 0.2164 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0763754 .0127317 6.00 0.000 .0514218 .1013291 avgcumdosew1 | -.5587715 .2733515 -2.04 0.041 -1.094531 -.0230124 bf4 | -.122574 .0262798 -4.66 0.000 -.1740815 -.0710665 _cons | -3.204746 .777467 -4.12 0.000 -4.728553 -1.680939 ------------------------------------------------------------------------------ 459 . 460 . 461 . 462 . di as input "female trimmed model for dose-home care impact in wv 1wave 1: do > se not signif" female trimmed model for dose-home care impact in wv 1wave 1: dose not signif 463 . di as input " female dose is not signif as main effect in dose - homecare imp > act " female dose is not signif as main effect in dose - homecare impact 464 . di as input " no female moderate interactions for dose-homecare impact" no female moderate interactions for dose-homecare impact 465 . 466 . scalar SigDoseHmcrFw1 = "yes" 467 . scalar MainEffhmcrFw1= "age b4 avgcumdosew1" 468 . scalar list MainEffhmcrFw1 = age b4 avgcumdosew1 hmcrMedFw1 = age avgcumdosew1 bf4 bf4m hmcrMedMw1 = age MainEffhmcrMw1 = age shjobw1 MainEffwkFw1 = age MainEffwkMw1 = age MainEffVactnMw1 = age radhlw1 VactnMedFw1 = age illw1 radhlw1 VactnMedMw1 = age VacatnModFw1 = none MainEffVactnFw1 = age radhlw1 bf7m SigDoseVactnFw1 = no VactnModMw1 = none inthobMedFw1 = age bf4 illw1 bf4m inthobMedMw1 = age inthobMw1 = age InthbModFw1 = none MainEffInthbFw1 = age radhlw1 bf4 SigdoseInthbFw1 = no InthbModMw1 = none MainEffInthbMw1 = age radhlw1 shfamw1 SigDoseInthbMw1 = no MainEffMw1 = radhlw1 bf4 bf40 sxlifeMedMw1 = radhlw1 sxlifeMedFw1 = age illw1 radhlw1 bf4 bf4m MainEffsxlifeFw1 = age bf4 bf4m SigDoseSxlifeFw1 = no SigDosesxlifeMw1 = no MainEffsxlifeMw1 = age bf4 bf40 SigDosePrbfmhmMw1 = no vactnModMw1 = none SigDoseVactnMw1 = no SxLifeModFw1 = no sxlifeModFw1 = none sxlifeModMw1 = none MaineffhmcrMw1 = age bf4 bf40 SigDoseMEhmcrW1 = no PrbfmhmMedFw1 = age bf4 PrbfmhmMedMw1 = age MainEffPrbfmhmFw1 = age radhlw1 bf4 MainEffPrbfmhmMw1 = age bf4 PrbfmhmModFw1 = none PrbfmhmModMw1 = none SigDosePrbfmhmFw1 = no SigDosePrbfhmMw1 = no MainEffPrbfhmMw1 = age bf4 MainEffVactnMw2 = age radhlw2 sxlifeMedMw2 = age illw2 SigDoseSxlifeFw2 = no MainEffsxlifeFw2 = age radhlw2 bf4 bf4m MainEffPrbsocMw2 = age radhlw2 shjobw2 MainEffhmcrFw2 = age hmcrMedFw2 = age bf4 MainEffwkFw2 = age MainEffwkMw2 = age MainEffPrbsocMw1 = age bf4m SigdoseMw1 = no ProbsocMedFw1 = age bf4 ProbsocMedMw1 = radhlw1 ProbsocModFw1 = none SigDoseProbsocMw1 = no hmcrmedMw1 = radhlw1 hmcrmedFw1 = age b4 b40 SigdosehmcrFw1 = no MainEffProbSocFw1 = age radhlw1 avgcumdosew1 shrelaw1 bf4 SigDoseProbsocFw1 = yes PrbsocModMw1 = shjobw1Xd1 shrelaw1Xd1 WkhmcrMw1 = age b4 WkModFw1 = ageXd1 hmcareMedFw1 = age illw1 hmcareMedMw1 = age SigDosehmcrFw1 = no wkMedMw1 = bf40 hmcrModFw1 = none SigDoseHmcrFw1 = yes WkMedMw1 = bf40 hmcrModMw1 = none SigDosehmcrMw1 = no wkMedFw1 = age b4 WKModMw1 = none SigDoseWkMw1 = no SigDoseWkFw1 = no SigDoseFw1 = no wkModFw1 = none wkModMw1 = none VactnMedMw2 = age inthobMedMw2 = age inthobMw2 = age PrbfmhmModMw2 = none MainEffProbSocFw2 = age radhlw2 avgcumdosew2 bf4 hmcrModMw2 = none MainEffhmcrMw2 = age wkMedFw2 = age b4 wkMedMw2 = age bf4 MainEffsxlifeMw2 = age bf4 bf40 shjobw2 shrelaw2 radhlw2 MainEffPrbfmhmMw2 = bf4 bf6 bf7 ProbsocMedFw2 = age bf4 radhlw2 hmcareMedFw2 = age bf4 WkhmcrMw2 = age b4 MainEffhmcrw2 = age hmcrModFw2 = none SigDoseHmcrFw2 = yes NumhmcrModMw2 = none SigDosehmcrMw2 = no SigdosehmcrFw2 = yes hmcrMedMw2 = age ageXillw2 SigDosehmcrFw2 = no MainEffhmcareMw2 = age WkMedMw2 = age ageXillw2 wkMedFw3 = radhlw3 age ageXillw3 bf40 bf4m bf1 VactnMedFw2 = age illw2 radhlw2 VacatnModFw2 = none MainEffVactnFw2 = age radhlw2 bf7m SigDoseVactnFw2 = no VactnModMw2 = none vactnModMw2 = none SigDoseVactnMw2 = no inthobMedFw2 = age bf4 illw2 bf4m InthbModFw2 = none MainEffInthbFw2 = age radhlw2 bf4 SigdoseInthbFw2 = no InthbModMw2 = none MainEffInthbMw2 = age radhlw2 shfamw2 SigDoseInthbMw2 = no MainEffMw2 = radhlw2 bf4 bf40 SigdoseMEinthob = no sxlifeMedFw2 = age illw2 radhlw2 bf4 bf4m SxLifeModFw2 = no sxlifeModFw2 = none sxlifeModMw2 = none SigDosesxlifeMw2 = no PrbfmhmMedFw2 = age bf4 PrbfmhmMedMw2 = age PrbfmhmModFw2 = none MainEffPrbfmhmFw2 = age bf4 bf40 SigDosePrbfmhmFw2 = no PrbfmhmModw2 = none SigDosePrbfmhmMw2 = no SigDosePrbfhmMw2 = no MainEffPrbfhmMw2 = bf4 bf6 bf7 ProbsocMedMw2 = age ProbsocModFw2 = none SigDoseProbsocFw2 = yes ProbSocModMw2 = none SigDoseProbsocMw2 = no PrbsocModMw2 = none SigdoseMw2 = none hmcareMedMw2 = age hmcareModFw2 = none MainEffhmcarew2 = age SigdoseHmcareFw2 = no hmcareModMw2 = none SigDoseHmcareMw2 = no NameMedMw2 = age ageXillw2 NumModMw2 = none SigDosehmcareMw2 = no SigDoseWKMw2 = no WkMedFw2 = age bf4 WkModFw2 = none WKModMw2 = none SigDoseWkMw2 = no SigDoseWkFw2 = no SigDoseFw2 = no wkModFw2 = none wkModMw2 = none VactnMedFw3 = age illw3 radhlw3 VactnMedMw3 = age illw3 VacatnModFw3 = none MainEffVactnFw3 = age radhlw3 deaw3 SigDoseVactnFw3 = no vactnModMw3 = none MainEffVactnMw3 = age bf7m radhlw3 SigDoseVactnMw3 = no sxLifeMedFw3 = age bf4 bf4m sxLifeMedMw3 = age illw3 InthbModFw3 = none MainEffInthbFw3 = age radhlw3 bf4 SigdoseInthbFw3 = no InthbMw3 = none MainEffInthbMw3 = age radhlw3 shfamw3 SigDoseInthbMw3 = no sxlifeMedFw3 = age illw3 radhlw3 bf4 bf4m sxlifeMedMw3 = age illw3 sxlifeModFw3 = none MainEffsxlifeFw3 = age radhlw3 bf4 bf4m shrelaw3 shfamw3 SigDoseSxlifeFw3 = no sxlifeModMw3 = none SigDosesxlifeMw3 = no MainEffsxlifeMw3 = age bf4 illw3 radhlw3 PrbfmhmMedFw3 = age bf4 PrbfmhmMedMw3 = age PrbfmhmModFw3 = none MainEffPrbfmhmFw3 = age bf4 bf40 SigDosePrbfmhmFw3 = no PrbfmhmModw3 = none SigDosePrbfmhmMw3 = no SigDosePrbfhmMw3 = no MainEffPrbfhmMw3 = bf1 bf4 dvcew3 bf7m ProbsocMedFw3 = age radhlw3 ProbsocMedMw3 = age ProbsocModFw3 = none MainEffProbSocFw3 = age radhlw3 illw3 Shrelaw3 avgcumodsew3 SigDoseProbsocFw3 = yes ProbSocModMw3 = none SigDoseProbsocMw3 = no MainEffPrbsocMw3 = age radhlw3 shjobw3 hmcareMedFw3 = age illw3 hmcareMedMw3 = age illw3 hmcareModFw3 = none SigdoseHmcareFw3 = no hmcareModMw3 = none MainEffhmcareMw3 = none SigDoseHmcareMw3 = no wkMedMw3 = bf8 age illw3 ageXillw3 wkModFw3 = none wkModMw3 = none MainEffwkFw3 = age MainEffwkMw3 = workM: age bf8 illw3 shjobw3 SigDoseWKMw3 = no SigDoseWkFw3 = no 469 . 470 . title4 "Fully trimmed Female Main effects model for dose=> homecare wave 1" ------------------------------------------------------------------------------- Fully trimmed Female Main effects model for dose=> homecare wave 1 ------------------------------------------------------------------------------- 471 . logit hp2hmcare age radhlw1 occ1w1-occ8w1 inc1w1-inc4w1 /// > avgcumdosew1 bf4 if gender==2 Iteration 0: log likelihood = -233.30573 Iteration 1: log likelihood = -177.44485 Iteration 2: log likelihood = -175.62926 Iteration 3: log likelihood = -175.62508 Iteration 4: log likelihood = -175.62508 Logistic regression Number of obs = 362 LR chi2(16) = 115.36 Prob > chi2 = 0.0000 Log likelihood = -175.62508 Pseudo R2 = 0.2472 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0961718 .0183122 5.25 0.000 .0602806 .132063 radhlw1 | -.0009397 .0037872 -0.25 0.804 -.0083624 .006483 occ1w1 | -1.536209 1.27581 -1.20 0.229 -4.036751 .9643328 occ2w1 | -1.066639 1.306018 -0.82 0.414 -3.626387 1.493109 occ3w1 | -.9187994 1.316697 -0.70 0.485 -3.499477 1.661879 occ4w1 | -1.809667 1.369371 -1.32 0.186 -4.493584 .8742504 occ5w1 | -2.474943 1.459934 -1.70 0.090 -5.336361 .3864749 occ6w1 | -2.384731 1.550879 -1.54 0.124 -5.424399 .6549365 occ7w1 | -.2736949 1.305569 -0.21 0.834 -2.832564 2.285174 occ8w1 | -.4828664 1.289472 -0.37 0.708 -3.010185 2.044453 inc1w1 | 1.237577 1.315563 0.94 0.347 -1.340878 3.816033 inc2w1 | 1.563876 1.263131 1.24 0.216 -.9118151 4.039568 inc3w1 | 1.858766 1.259322 1.48 0.140 -.6094602 4.326992 inc4w1 | 1.650938 1.369374 1.21 0.228 -1.032987 4.334862 avgcumdosew1 | -.6169468 .2918407 -2.11 0.035 -1.188944 -.0449496 bf4 | -.1423744 .0286191 -4.97 0.000 -.1984668 -.086282 _cons | -4.384177 1.11318 -3.94 0.000 -6.56597 -2.202384 ------------------------------------------------------------------------------ 472 . 473 . title4 "Super trimmed Female Main effects model for dose=> homecare wave 1" ------------------------------------------------------------------------------- Super trimmed Female Main effects model for dose=> homecare wave 1 ------------------------------------------------------------------------------- 474 . sw, pr(.1): logit hp2hmcare age radhlw1 occ1w1-occ8w1 inc1w1-inc4w1 /// > avgcumdosew1 bf4 bf7 if gender==2 begin with full model p = 0.8548 >= 0.1000 removing occ7w1 p = 0.8444 >= 0.1000 removing radhlw1 p = 0.7229 >= 0.1000 removing occ8w1 p = 0.3325 >= 0.1000 removing occ3w1 p = 0.3694 >= 0.1000 removing occ2w1 p = 0.2150 >= 0.1000 removing inc1w1 p = 0.2936 >= 0.1000 removing inc4w1 p = 0.2169 >= 0.1000 removing inc2w1 p = 0.2571 >= 0.1000 removing occ4w1 p = 0.1614 >= 0.1000 removing occ1w1 p = 0.2137 >= 0.1000 removing inc3w1 p = 0.2479 >= 0.1000 removing occ6w1 p = 0.1543 >= 0.1000 removing occ5w1 p = 0.1037 >= 0.1000 removing bf7 Logistic regression Number of obs = 362 LR chi2(3) = 100.40 Prob > chi2 = 0.0000 Log likelihood = -183.1063 Pseudo R2 = 0.2152 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0762363 .0127384 5.98 0.000 .0512695 .1012031 bf4 | -.1224238 .0262785 -4.66 0.000 -.1739286 -.0709189 avgcumdosew1 | -.5588394 .2733844 -2.04 0.041 -1.094663 -.0230158 _cons | -3.198013 .7776298 -4.11 0.000 -4.72214 -1.673887 ------------------------------------------------------------------------------ 475 . 476 . cap gen radhlw1Xd1 = radhlw1*avgcumdosew1 477 . 478 . 479 . title4 "female main effect plus interaction model" ------------------------------------------------------------------------------- female main effect plus interaction model ------------------------------------------------------------------------------- 480 . xi:logit hp2hmcare age avgcumdosew1 bf4 ageXd1 bf4Xd1 if gender==2 Iteration 0: log likelihood = -233.72859 Iteration 1: log likelihood = -183.82966 Iteration 2: log likelihood = -182.5885 Iteration 3: log likelihood = -182.58185 Iteration 4: log likelihood = -182.58185 Logistic regression Number of obs = 363 LR chi2(5) = 102.29 Prob > chi2 = 0.0000 Log likelihood = -182.58185 Pseudo R2 = 0.2188 ------------------------------------------------------------------------------ hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0849926 .0157492 5.40 0.000 .0541247 .1158604 avgcumdosew1 | 1.600207 2.155971 0.74 0.458 -2.625418 5.825832 bf4 | -.1093288 .034593 -3.16 0.002 -.1771298 -.0415278 ageXd1 | -.0339202 .0371133 -0.91 0.361 -.1066609 .0388205 bf4Xd1 | -.0454367 .0710033 -0.64 0.522 -.1846007 .0937273 _cons | -3.755837 .945655 -3.97 0.000 -5.609287 -1.902388 ------------------------------------------------------------------------------ 481 . 482 . * there are no significant moderators for female dose=hmcare relationship 483 . scalar hmcrModFw1 = "none" 484 . 485 . 486 . 487 . scalar SigdosehmcrFw1="no" 488 . 489 . 490 . scalar MainEffhmcrFw1 = "age bf4" 491 . 492 . 493 . scalar hmcrModFw1 = "none" 494 . 495 . 496 . title4 "Mediator relationships for home care are tested below" ------------------------------------------------------------------------------- Mediator relationships for home care are tested below ------------------------------------------------------------------------------- 497 . title4 "H1 pt 2 wave 1 Mediation of home care testing for males" ------------------------------------------------------------------------------- H1 pt 2 wave 1 Mediation of home care testing for males ------------------------------------------------------------------------------- 498 . 499 . 500 . 501 . correlate bf4 age if gender==1 (obs=340) | bf4 age -------------+------------------ bf4 | 1.0000 age | -0.4041 1.0000 502 . 503 . des bf4 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) 504 . 505 . glm bf4 avgcumdosew1 if gender==1, fam(gauss) link(identity) Iteration 0: log likelihood = -1026.9659 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 24.75428 Deviance = 8366.946191 (1/df) Deviance = 24.75428 Pearson = 8366.946191 (1/df) Pearson = 24.75428 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.05274 Log likelihood = -1026.965868 BIC = 6396.763 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.1031788 .161925 -0.64 0.524 -.4205461 .2141884 _cons | 12.54134 .2786337 45.01 0.000 11.99523 13.08746 ------------------------------------------------------------------------------ 506 . glm hp2hmcare bf4 if gender==1, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 268.3422 Iteration 2: deviance = 265.0048 Iteration 3: deviance = 264.9272 Iteration 4: deviance = 264.9269 Iteration 5: deviance = 264.9269 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 264.9268544 (1/df) Deviance = .7838073 Pearson = 286.2423079 (1/df) Pearson = .8468707 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1705.257 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1429587 .015032 -9.51 0.000 -.1724209 -.1134964 _cons | .7855991 .1813087 4.33 0.000 .4302406 1.140958 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 507 . 508 . 509 . glm age avgcumdosew1 if gender==1, fam(gauss) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 510 . glm hp2hmcare age if gender==1, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 320.8295 Iteration 2: deviance = 320.0337 Iteration 3: deviance = 320.0328 Iteration 4: deviance = 320.0328 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 320.0327771 (1/df) Deviance = .9468425 Pearson = 341.699231 (1/df) Pearson = 1.010944 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1650.151 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .032545 .0064212 5.07 0.000 .0199596 .0451304 _cons | -2.483665 .3435565 -7.23 0.000 -3.157023 -1.810306 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 511 . 512 . scalar hmcrMedMw1 = "radhlw1" 513 . 514 . 515 . title4 "radhlw1 is a possible mediator for men of hmcare in wave 1" ------------------------------------------------------------------------------- radhlw1 is a possible mediator for men of hmcare in wave 1 ------------------------------------------------------------------------------- 516 . glm radhlw1 avgcumdosew1 if gender==1, fam(gaussian) link(identity) Iteration 0: log likelihood = -1710.3417 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1378.645 Deviance = 465981.8893 (1/df) Deviance = 1378.645 Pearson = 465981.8893 (1/df) Pearson = 1378.645 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.0726 Log likelihood = -1710.341694 BIC = 464011.7 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 2.398285 1.208412 1.98 0.047 .0298407 4.766729 _cons | 44.66477 2.079384 21.48 0.000 40.58925 48.74029 ------------------------------------------------------------------------------ 517 . glm hp2hmcare radhlw1 if gender==1 , fam(binomial) irls scale(dev) link(probi > t) Iteration 1: deviance = 330.2185 Iteration 2: deviance = 329.7978 Iteration 3: deviance = 329.7977 Iteration 4: deviance = 329.7977 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 329.797656 (1/df) Deviance = .9757327 Pearson = 341.5925114 (1/df) Pearson = 1.010629 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1640.386 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .008257 .0020717 3.99 0.000 .0041965 .0123175 _cons | -1.236257 .1340392 -9.22 0.000 -1.498968 -.9735446 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 518 . 519 . title4 "Testing possible female home care mediators for home care in wave 1" ------------------------------------------------------------------------------- Testing possible female home care mediators for home care in wave 1 ------------------------------------------------------------------------------- 520 . 521 . title4 "Test of age as possible female mediator in wave 1" ------------------------------------------------------------------------------- Test of age as possible female mediator in wave 1 ------------------------------------------------------------------------------- 522 . glm age avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 523 . glm hp2hmcare age if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 393.7958 Iteration 2: deviance = 393.6955 Iteration 3: deviance = 393.6955 Iteration 4: deviance = 393.6955 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 393.6954976 (1/df) Deviance = 1.090569 Pearson = 375.4421438 (1/df) Pearson = 1.040006 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1734.184 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0533501 .0069587 7.67 0.000 .0397113 .0669889 _cons | -3.144653 .3710751 -8.47 0.000 -3.871946 -2.417359 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 524 . title4 "age is a possible mediator for women" ------------------------------------------------------------------------------- age is a possible mediator for women ------------------------------------------------------------------------------- 525 . 526 . 527 . title4 "Test of b40 as female mediator of home care in wave 1" ------------------------------------------------------------------------------- Test of b40 as female mediator of home care in wave 1 ------------------------------------------------------------------------------- 528 . glm bf40 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -818.49948 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 5.351018 Deviance = 1931.717442 (1/df) Deviance = 5.351018 Pearson = 1931.717442 (1/df) Pearson = 5.351018 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 4.520658 Log likelihood = -818.4994757 BIC = -196.162 ------------------------------------------------------------------------------ | OIM bf40 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .4524905 .2213302 2.04 0.041 .0186913 .8862898 _cons | 3.013471 .1423225 21.17 0.000 2.734524 3.292417 ------------------------------------------------------------------------------ 529 . glm hp2hmcare bf40 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 462.764 Iteration 2: deviance = 462.2054 Iteration 3: deviance = 462.2053 Iteration 4: deviance = 462.2053 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 462.2053395 (1/df) Deviance = 1.280347 Pearson = 362.3613268 (1/df) Pearson = 1.003771 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1665.674 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf40 | .0666058 .0327979 2.03 0.042 .0023231 .1308886 _cons | -.6159451 .1313241 -4.69 0.000 -.8733356 -.3585546 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 530 . title4 "b40 with age is a mediator for women" ------------------------------------------------------------------------------- b40 with age is a mediator for women ------------------------------------------------------------------------------- 531 . 532 . 533 . 534 . title4 "bf4 Test of female mediation of home care in wave 1" ------------------------------------------------------------------------------- bf4 Test of female mediation of home care in wave 1 ------------------------------------------------------------------------------- 535 . glm bf4 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 536 . glm hp2hmcare bf4 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 410.1424 Iteration 2: deviance = 410.1166 Iteration 3: deviance = 410.1166 Iteration 4: deviance = 410.1166 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 410.1166157 (1/df) Deviance = 1.136057 Pearson = 357.1140873 (1/df) Pearson = .9892357 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1717.763 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1024685 .0149808 -6.84 0.000 -.1318304 -.0731067 _cons | .6214741 .1665648 3.73 0.000 .295013 .9479351 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 537 . title4 "bf4 alone is a mediator for women" ------------------------------------------------------------------------------- bf4 alone is a mediator for women ------------------------------------------------------------------------------- 538 . 539 . 540 . glm illw1 avgcumdosew1 if gender==2, fam(gauss) link(identity) Iteration 0: log likelihood = -259.70777 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = .2462383 Deviance = 88.89203958 (1/df) Deviance = .2462383 Pearson = 88.89203958 (1/df) Pearson = .2462383 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 1.441916 Log likelihood = -259.7077741 BIC = -2038.987 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0275426 .1585711 _cons | .1570822 .0305304 5.15 0.000 .0972436 .2169207 ------------------------------------------------------------------------------ 541 . glm hp2hmcare illw1 if gender==2, fam(binomial) link(probit) irls scale(dev) Iteration 1: deviance = 451.3594 Iteration 2: deviance = 450.782 Iteration 3: deviance = 450.7819 Iteration 4: deviance = 450.7819 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 450.7819203 (1/df) Deviance = 1.248703 Pearson = 362.4981549 (1/df) Pearson = 1.00415 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1677.098 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .5711151 .1643397 3.48 0.001 .2490151 .8932151 _cons | -.5078804 .0821368 -6.18 0.000 -.6688655 -.3468953 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 542 . 543 . 544 . title4 "multivariate Test of female mediation of home care in wave 1" ------------------------------------------------------------------------------- multivariate Test of female mediation of home care in wave 1 ------------------------------------------------------------------------------- 545 . mvreg bf4 age bf40 = avgcumdosew1 if gender==2 Equation Obs Parms RMSE "R-sq" F P ---------------------------------------------------------------------- bf4 363 2 5.149837 0.0253 9.376729 0.0024 age 363 2 11.6814 0.0338 12.64139 0.0004 bf40 363 2 2.313227 0.0114 4.179627 0.0416 ------------------------------------------------------------------------------ | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.477832 -.5398375 _cons | 10.99384 .3168463 34.70 0.000 10.37075 11.61694 -------------+---------------------------------------------------------------- age | avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.775898 6.171859 _cons | 48.88157 .7187038 68.01 0.000 47.4682 50.29494 -------------+---------------------------------------------------------------- bf40 | avgcumdosew1 | .4524905 .2213302 2.04 0.042 .017232 .887749 _cons | 3.013471 .1423225 21.17 0.000 2.733585 3.293356 ------------------------------------------------------------------------------ 546 . glm hp2hmcare bf4 age bf40 if gender==2, fam(binomial) link(probit) irls /// > scale(dev) Iteration 1: deviance = 373.2413 Iteration 2: deviance = 372.8437 Iteration 3: deviance = 372.8436 Iteration 4: deviance = 372.8436 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 359 (IRLS EIM) Scale parameter = 1 Deviance = 372.8435707 (1/df) Deviance = 1.038561 Pearson = 384.2294832 (1/df) Pearson = 1.070277 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1743.247 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.0701088 .0159082 -4.41 0.000 -.1012882 -.0389293 age | .0419129 .007264 5.77 0.000 .0276756 .0561501 bf40 | -.0040389 .0340124 -0.12 0.905 -.070702 .0626242 _cons | -1.841186 .4704258 -3.91 0.000 -2.763204 -.9191686 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 547 . title4 "When bf4 & age are together" /// > "only age & b4 are a wave 1 mediators for women" ------------------------------------------------------------------------------- When bf4 & age are together ------------------------------------------------------------------------------- 548 . 549 . glm radhlw1 avgcumdosew1 if gender==2, fam(gaussian) link(identity) Iteration 0: log likelihood = -1821.9477 Generalized linear models No. of obs = 362 Optimization : ML Residual df = 360 Scale parameter = 1385.301 Deviance = 498708.3025 (1/df) Deviance = 1385.301 Pearson = 498708.3025 (1/df) Pearson = 1385.301 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.07706 Log likelihood = -1821.947718 BIC = 496587.3 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.789972 3.562254 1.06 0.287 -3.191917 10.77186 _cons | 55.44948 2.293757 24.17 0.000 50.9538 59.94516 ------------------------------------------------------------------------------ 550 . glm hp2hmcare radhlw1 if gender==2 , fam(binomial) irls scale(dev) link(probi > t) Iteration 1: deviance = 463.5751 Iteration 2: deviance = 463.0208 Iteration 3: deviance = 463.0207 Iteration 4: deviance = 463.0207 Generalized linear models No. of obs = 362 Optimization : MQL Fisher scoring Residual df = 360 (IRLS EIM) Scale parameter = 1 Deviance = 463.0207118 (1/df) Deviance = 1.286169 Pearson = 361.7039664 (1/df) Pearson = 1.004733 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1657.971 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0034921 .0020846 1.68 0.094 -.0005936 .0075779 _cons | -.5996677 .1437134 -4.17 0.000 -.8813409 -.3179945 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 551 . 552 . title "Test of female mediation of home care in wave 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Test of female mediation of home care in wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:14:13 ***** ******************************************************************************* ******************************************************************************* 553 . mvreg bf4 age bf40 illw1 = avgcumdosew1 if gender==2 Equation Obs Parms RMSE "R-sq" F P ---------------------------------------------------------------------- bf4 363 2 5.149837 0.0253 9.376729 0.0024 age 363 2 11.6814 0.0338 12.64139 0.0004 bf40 363 2 2.313227 0.0114 4.179627 0.0416 illw1 363 2 .4962241 0.0052 1.904017 0.1685 ------------------------------------------------------------------------------ | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.477832 -.5398375 _cons | 10.99384 .3168463 34.70 0.000 10.37075 11.61694 -------------+---------------------------------------------------------------- age | avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.775898 6.171859 _cons | 48.88157 .7187038 68.01 0.000 47.4682 50.29494 -------------+---------------------------------------------------------------- bf40 | avgcumdosew1 | .4524905 .2213302 2.04 0.042 .017232 .887749 _cons | 3.013471 .1423225 21.17 0.000 2.733585 3.293356 -------------+---------------------------------------------------------------- illw1 | avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0278557 .1588841 _cons | .1570822 .0305304 5.15 0.000 .0970423 .217122 ------------------------------------------------------------------------------ 554 . glm hp2hmcare bf4 age bf40 illw1 if gender==2, fam(binomial) /// > link(probit) irls scale(dev) Iteration 1: deviance = 368.0172 Iteration 2: deviance = 367.5924 Iteration 3: deviance = 367.5922 Iteration 4: deviance = 367.5922 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 358 (IRLS EIM) Scale parameter = 1 Deviance = 367.5922256 (1/df) Deviance = 1.026794 Pearson = 388.4520039 (1/df) Pearson = 1.085061 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1742.604 ------------------------------------------------------------------------------ | EIM hp2hmcare | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.0651744 .01601 -4.07 0.000 -.0965533 -.0337954 age | .0411498 .007244 5.68 0.000 .0269518 .0553477 bf40 | -.0137692 .0343692 -0.40 0.689 -.0811316 .0535933 illw1 | .3518084 .15798 2.23 0.026 .0421733 .6614435 _cons | -1.887052 .4701989 -4.01 0.000 -2.808625 -.9654796 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 555 . qui: { When all are together only age and b4 are a wave 1 mediators for women 556 . 557 . 558 . scalar hmcrMedFw1 = "age bf4" 559 . 560 . scalar SigDosehmcrMw1 = "no" 561 . scalar MainEffhmcrMw1 = "age" 562 . scalar MainEffhmcrFw1 = "age" 563 . scalar hmcrmedFw1 = "age b4 b40" 564 . scalar hmcrmedMw1 = "radhlw1" 565 . 566 . scalar hmcareMedMw1 = "age " 567 . scalar hmcareMedFw1 = "age illw1" 568 . set more off 569 . scalar list MainEffhmcrFw1 = age MainEffhmcrMw1 = age hmcrMedFw1 = age bf4 hmcrMedMw1 = radhlw1 MainEffwkFw1 = age MainEffwkMw1 = age MainEffVactnMw1 = age radhlw1 VactnMedFw1 = age illw1 radhlw1 VactnMedMw1 = age VacatnModFw1 = none MainEffVactnFw1 = age radhlw1 bf7m SigDoseVactnFw1 = no VactnModMw1 = none inthobMedFw1 = age bf4 illw1 bf4m inthobMedMw1 = age inthobMw1 = age InthbModFw1 = none MainEffInthbFw1 = age radhlw1 bf4 SigdoseInthbFw1 = no InthbModMw1 = none MainEffInthbMw1 = age radhlw1 shfamw1 SigDoseInthbMw1 = no MainEffMw1 = radhlw1 bf4 bf40 sxlifeMedMw1 = radhlw1 sxlifeMedFw1 = age illw1 radhlw1 bf4 bf4m MainEffsxlifeFw1 = age bf4 bf4m SigDoseSxlifeFw1 = no SigDosesxlifeMw1 = no MainEffsxlifeMw1 = age bf4 bf40 SigDosePrbfmhmMw1 = no vactnModMw1 = none SigDoseVactnMw1 = no SxLifeModFw1 = no sxlifeModFw1 = none sxlifeModMw1 = none MaineffhmcrMw1 = age bf4 bf40 SigDoseMEhmcrW1 = no PrbfmhmMedFw1 = age bf4 PrbfmhmMedMw1 = age MainEffPrbfmhmFw1 = age radhlw1 bf4 MainEffPrbfmhmMw1 = age bf4 PrbfmhmModFw1 = none PrbfmhmModMw1 = none SigDosePrbfmhmFw1 = no SigDosePrbfhmMw1 = no MainEffPrbfhmMw1 = age bf4 MainEffVactnMw2 = age radhlw2 sxlifeMedMw2 = age illw2 SigDoseSxlifeFw2 = no MainEffsxlifeFw2 = age radhlw2 bf4 bf4m MainEffPrbsocMw2 = age radhlw2 shjobw2 MainEffhmcrFw2 = age hmcrMedFw2 = age bf4 MainEffwkFw2 = age MainEffwkMw2 = age MainEffPrbsocMw1 = age bf4m SigdoseMw1 = no ProbsocMedFw1 = age bf4 ProbsocMedMw1 = radhlw1 ProbsocModFw1 = none SigDoseProbsocMw1 = no hmcrmedMw1 = radhlw1 hmcrmedFw1 = age b4 b40 SigdosehmcrFw1 = no MainEffProbSocFw1 = age radhlw1 avgcumdosew1 shrelaw1 bf4 SigDoseProbsocFw1 = yes PrbsocModMw1 = shjobw1Xd1 shrelaw1Xd1 WkhmcrMw1 = age b4 WkModFw1 = ageXd1 hmcareMedFw1 = age illw1 hmcareMedMw1 = age SigDosehmcrFw1 = no wkMedMw1 = bf40 hmcrModFw1 = none SigDoseHmcrFw1 = yes WkMedMw1 = bf40 hmcrModMw1 = none SigDosehmcrMw1 = no wkMedFw1 = age b4 WKModMw1 = none SigDoseWkMw1 = no SigDoseWkFw1 = no SigDoseFw1 = no wkModFw1 = none wkModMw1 = none VactnMedMw2 = age inthobMedMw2 = age inthobMw2 = age PrbfmhmModMw2 = none MainEffProbSocFw2 = age radhlw2 avgcumdosew2 bf4 hmcrModMw2 = none MainEffhmcrMw2 = age wkMedFw2 = age b4 wkMedMw2 = age bf4 MainEffsxlifeMw2 = age bf4 bf40 shjobw2 shrelaw2 radhlw2 MainEffPrbfmhmMw2 = bf4 bf6 bf7 ProbsocMedFw2 = age bf4 radhlw2 hmcareMedFw2 = age bf4 WkhmcrMw2 = age b4 MainEffhmcrw2 = age hmcrModFw2 = none SigDoseHmcrFw2 = yes NumhmcrModMw2 = none SigDosehmcrMw2 = no SigdosehmcrFw2 = yes hmcrMedMw2 = age ageXillw2 SigDosehmcrFw2 = no MainEffhmcareMw2 = age WkMedMw2 = age ageXillw2 wkMedFw3 = radhlw3 age ageXillw3 bf40 bf4m bf1 VactnMedFw2 = age illw2 radhlw2 VacatnModFw2 = none MainEffVactnFw2 = age radhlw2 bf7m SigDoseVactnFw2 = no VactnModMw2 = none vactnModMw2 = none SigDoseVactnMw2 = no inthobMedFw2 = age bf4 illw2 bf4m InthbModFw2 = none MainEffInthbFw2 = age radhlw2 bf4 SigdoseInthbFw2 = no InthbModMw2 = none MainEffInthbMw2 = age radhlw2 shfamw2 SigDoseInthbMw2 = no MainEffMw2 = radhlw2 bf4 bf40 SigdoseMEinthob = no sxlifeMedFw2 = age illw2 radhlw2 bf4 bf4m SxLifeModFw2 = no sxlifeModFw2 = none sxlifeModMw2 = none SigDosesxlifeMw2 = no PrbfmhmMedFw2 = age bf4 PrbfmhmMedMw2 = age PrbfmhmModFw2 = none MainEffPrbfmhmFw2 = age bf4 bf40 SigDosePrbfmhmFw2 = no PrbfmhmModw2 = none SigDosePrbfmhmMw2 = no SigDosePrbfhmMw2 = no MainEffPrbfhmMw2 = bf4 bf6 bf7 ProbsocMedMw2 = age ProbsocModFw2 = none SigDoseProbsocFw2 = yes ProbSocModMw2 = none SigDoseProbsocMw2 = no PrbsocModMw2 = none SigdoseMw2 = none hmcareMedMw2 = age hmcareModFw2 = none MainEffhmcarew2 = age SigdoseHmcareFw2 = no hmcareModMw2 = none SigDoseHmcareMw2 = no NameMedMw2 = age ageXillw2 NumModMw2 = none SigDosehmcareMw2 = no SigDoseWKMw2 = no WkMedFw2 = age bf4 WkModFw2 = none WKModMw2 = none SigDoseWkMw2 = no SigDoseWkFw2 = no SigDoseFw2 = no wkModFw2 = none wkModMw2 = none VactnMedFw3 = age illw3 radhlw3 VactnMedMw3 = age illw3 VacatnModFw3 = none MainEffVactnFw3 = age radhlw3 deaw3 SigDoseVactnFw3 = no vactnModMw3 = none MainEffVactnMw3 = age bf7m radhlw3 SigDoseVactnMw3 = no sxLifeMedFw3 = age bf4 bf4m sxLifeMedMw3 = age illw3 InthbModFw3 = none MainEffInthbFw3 = age radhlw3 bf4 SigdoseInthbFw3 = no InthbMw3 = none MainEffInthbMw3 = age radhlw3 shfamw3 SigDoseInthbMw3 = no sxlifeMedFw3 = age illw3 radhlw3 bf4 bf4m sxlifeMedMw3 = age illw3 sxlifeModFw3 = none MainEffsxlifeFw3 = age radhlw3 bf4 bf4m shrelaw3 shfamw3 SigDoseSxlifeFw3 = no sxlifeModMw3 = none SigDosesxlifeMw3 = no MainEffsxlifeMw3 = age bf4 illw3 radhlw3 PrbfmhmMedFw3 = age bf4 PrbfmhmMedMw3 = age PrbfmhmModFw3 = none MainEffPrbfmhmFw3 = age bf4 bf40 SigDosePrbfmhmFw3 = no PrbfmhmModw3 = none SigDosePrbfmhmMw3 = no SigDosePrbfhmMw3 = no MainEffPrbfhmMw3 = bf1 bf4 dvcew3 bf7m ProbsocMedFw3 = age radhlw3 ProbsocMedMw3 = age ProbsocModFw3 = none MainEffProbSocFw3 = age radhlw3 illw3 Shrelaw3 avgcumodsew3 SigDoseProbsocFw3 = yes ProbSocModMw3 = none SigDoseProbsocMw3 = no MainEffPrbsocMw3 = age radhlw3 shjobw3 hmcareMedFw3 = age illw3 hmcareMedMw3 = age illw3 hmcareModFw3 = none SigdoseHmcareFw3 = no hmcareModMw3 = none MainEffhmcareMw3 = none SigDoseHmcareMw3 = no wkMedMw3 = bf8 age illw3 ageXillw3 wkModFw3 = none wkModMw3 = none MainEffwkFw3 = age MainEffwkMw3 = workM: age bf8 illw3 shjobw3 SigDoseWKMw3 = no SigDoseWkFw3 = no 570 . 571 . * conclusion "age & illw1 are main effects as possible male & female mediator > s" 572 . * conclusion title "their interaction is not a mediator" 573 . 574 . title4 "2. summary matrix construction for H1 pt 2 wave 1 dose=>Home care imp > act" ------------------------------------------------------------------------------- 2. summary matrix construction for H1 pt 2 wave 1 dose=>Home care impact ------------------------------------------------------------------------------- 575 . set more off 576 . matrix define hmcrMw1 = J(1,8, 0) 577 . matrix define hmcrFw1 = J(1,8, 0) 578 . matrix colnames hmcrMw1= hypnum ptnum wave gender medsig numMAsig numModsig > /// > numMed 579 . matrix colnames hmcrFw1= hypnum ptnum wave gender medsig numMAsig numModsig > /// > numMed 580 . matrix rownames hmcrMw1 = hmcareM 581 . matrix rownames hmcrFw1 = hmcareF 582 . matrix define hmcrMw1= (1, 2, 1, 1, 0 , 2, 0 , 1 ) 583 . matrix define hmcrFw1= (1, 2, 1, 2, 0 ,2, 0 , 2 ) 584 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1 \ hmcrMw1 \ hmcrFw1) 585 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 586 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 587 . matrix rownames H1pt2w1 = wkMw1 wkFw1 hmcrMw1 hmcrFw1 588 . matlist H1pt2w1 | hypnum ptnum wave gender medsig numMAsig > numModsig numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 hmcrMw1 | 1 2 1 1 0 2 > 0 1 hmcrFw1 | 1 2 1 2 0 2 > 0 2 589 . 590 . * see scalar list for names of variables 591 . scalar list MainEffhmcrFw1 = age MainEffhmcrMw1 = age hmcrMedFw1 = age bf4 hmcrMedMw1 = radhlw1 MainEffwkFw1 = age MainEffwkMw1 = age MainEffVactnMw1 = age radhlw1 VactnMedFw1 = age illw1 radhlw1 VactnMedMw1 = age VacatnModFw1 = none MainEffVactnFw1 = age radhlw1 bf7m SigDoseVactnFw1 = no VactnModMw1 = none inthobMedFw1 = age bf4 illw1 bf4m inthobMedMw1 = age inthobMw1 = age InthbModFw1 = none MainEffInthbFw1 = age radhlw1 bf4 SigdoseInthbFw1 = no InthbModMw1 = none MainEffInthbMw1 = age radhlw1 shfamw1 SigDoseInthbMw1 = no MainEffMw1 = radhlw1 bf4 bf40 sxlifeMedMw1 = radhlw1 sxlifeMedFw1 = age illw1 radhlw1 bf4 bf4m MainEffsxlifeFw1 = age bf4 bf4m SigDoseSxlifeFw1 = no SigDosesxlifeMw1 = no MainEffsxlifeMw1 = age bf4 bf40 SigDosePrbfmhmMw1 = no vactnModMw1 = none SigDoseVactnMw1 = no SxLifeModFw1 = no sxlifeModFw1 = none sxlifeModMw1 = none MaineffhmcrMw1 = age bf4 bf40 SigDoseMEhmcrW1 = no PrbfmhmMedFw1 = age bf4 PrbfmhmMedMw1 = age MainEffPrbfmhmFw1 = age radhlw1 bf4 MainEffPrbfmhmMw1 = age bf4 PrbfmhmModFw1 = none PrbfmhmModMw1 = none SigDosePrbfmhmFw1 = no SigDosePrbfhmMw1 = no MainEffPrbfhmMw1 = age bf4 MainEffVactnMw2 = age radhlw2 sxlifeMedMw2 = age illw2 SigDoseSxlifeFw2 = no MainEffsxlifeFw2 = age radhlw2 bf4 bf4m MainEffPrbsocMw2 = age radhlw2 shjobw2 MainEffhmcrFw2 = age hmcrMedFw2 = age bf4 MainEffwkFw2 = age MainEffwkMw2 = age MainEffPrbsocMw1 = age bf4m SigdoseMw1 = no ProbsocMedFw1 = age bf4 ProbsocMedMw1 = radhlw1 ProbsocModFw1 = none SigDoseProbsocMw1 = no hmcrmedMw1 = radhlw1 hmcrmedFw1 = age b4 b40 SigdosehmcrFw1 = no MainEffProbSocFw1 = age radhlw1 avgcumdosew1 shrelaw1 bf4 SigDoseProbsocFw1 = yes PrbsocModMw1 = shjobw1Xd1 shrelaw1Xd1 WkhmcrMw1 = age b4 WkModFw1 = ageXd1 hmcareMedFw1 = age illw1 hmcareMedMw1 = age SigDosehmcrFw1 = no wkMedMw1 = bf40 hmcrModFw1 = none SigDoseHmcrFw1 = yes WkMedMw1 = bf40 hmcrModMw1 = none SigDosehmcrMw1 = no wkMedFw1 = age b4 WKModMw1 = none SigDoseWkMw1 = no SigDoseWkFw1 = no SigDoseFw1 = no wkModFw1 = none wkModMw1 = none VactnMedMw2 = age inthobMedMw2 = age inthobMw2 = age PrbfmhmModMw2 = none MainEffProbSocFw2 = age radhlw2 avgcumdosew2 bf4 hmcrModMw2 = none MainEffhmcrMw2 = age wkMedFw2 = age b4 wkMedMw2 = age bf4 MainEffsxlifeMw2 = age bf4 bf40 shjobw2 shrelaw2 radhlw2 MainEffPrbfmhmMw2 = bf4 bf6 bf7 ProbsocMedFw2 = age bf4 radhlw2 hmcareMedFw2 = age bf4 WkhmcrMw2 = age b4 MainEffhmcrw2 = age hmcrModFw2 = none SigDoseHmcrFw2 = yes NumhmcrModMw2 = none SigDosehmcrMw2 = no SigdosehmcrFw2 = yes hmcrMedMw2 = age ageXillw2 SigDosehmcrFw2 = no MainEffhmcareMw2 = age WkMedMw2 = age ageXillw2 wkMedFw3 = radhlw3 age ageXillw3 bf40 bf4m bf1 VactnMedFw2 = age illw2 radhlw2 VacatnModFw2 = none MainEffVactnFw2 = age radhlw2 bf7m SigDoseVactnFw2 = no VactnModMw2 = none vactnModMw2 = none SigDoseVactnMw2 = no inthobMedFw2 = age bf4 illw2 bf4m InthbModFw2 = none MainEffInthbFw2 = age radhlw2 bf4 SigdoseInthbFw2 = no InthbModMw2 = none MainEffInthbMw2 = age radhlw2 shfamw2 SigDoseInthbMw2 = no MainEffMw2 = radhlw2 bf4 bf40 SigdoseMEinthob = no sxlifeMedFw2 = age illw2 radhlw2 bf4 bf4m SxLifeModFw2 = no sxlifeModFw2 = none sxlifeModMw2 = none SigDosesxlifeMw2 = no PrbfmhmMedFw2 = age bf4 PrbfmhmMedMw2 = age PrbfmhmModFw2 = none MainEffPrbfmhmFw2 = age bf4 bf40 SigDosePrbfmhmFw2 = no PrbfmhmModw2 = none SigDosePrbfmhmMw2 = no SigDosePrbfhmMw2 = no MainEffPrbfhmMw2 = bf4 bf6 bf7 ProbsocMedMw2 = age ProbsocModFw2 = none SigDoseProbsocFw2 = yes ProbSocModMw2 = none SigDoseProbsocMw2 = no PrbsocModMw2 = none SigdoseMw2 = none hmcareMedMw2 = age hmcareModFw2 = none MainEffhmcarew2 = age SigdoseHmcareFw2 = no hmcareModMw2 = none SigDoseHmcareMw2 = no NameMedMw2 = age ageXillw2 NumModMw2 = none SigDosehmcareMw2 = no SigDoseWKMw2 = no WkMedFw2 = age bf4 WkModFw2 = none WKModMw2 = none SigDoseWkMw2 = no SigDoseWkFw2 = no SigDoseFw2 = no wkModFw2 = none wkModMw2 = none VactnMedFw3 = age illw3 radhlw3 VactnMedMw3 = age illw3 VacatnModFw3 = none MainEffVactnFw3 = age radhlw3 deaw3 SigDoseVactnFw3 = no vactnModMw3 = none MainEffVactnMw3 = age bf7m radhlw3 SigDoseVactnMw3 = no sxLifeMedFw3 = age bf4 bf4m sxLifeMedMw3 = age illw3 InthbModFw3 = none MainEffInthbFw3 = age radhlw3 bf4 SigdoseInthbFw3 = no InthbMw3 = none MainEffInthbMw3 = age radhlw3 shfamw3 SigDoseInthbMw3 = no sxlifeMedFw3 = age illw3 radhlw3 bf4 bf4m sxlifeMedMw3 = age illw3 sxlifeModFw3 = none MainEffsxlifeFw3 = age radhlw3 bf4 bf4m shrelaw3 shfamw3 SigDoseSxlifeFw3 = no sxlifeModMw3 = none SigDosesxlifeMw3 = no MainEffsxlifeMw3 = age bf4 illw3 radhlw3 PrbfmhmMedFw3 = age bf4 PrbfmhmMedMw3 = age PrbfmhmModFw3 = none MainEffPrbfmhmFw3 = age bf4 bf40 SigDosePrbfmhmFw3 = no PrbfmhmModw3 = none SigDosePrbfmhmMw3 = no SigDosePrbfhmMw3 = no MainEffPrbfhmMw3 = bf1 bf4 dvcew3 bf7m ProbsocMedFw3 = age radhlw3 ProbsocMedMw3 = age ProbsocModFw3 = none MainEffProbSocFw3 = age radhlw3 illw3 Shrelaw3 avgcumodsew3 SigDoseProbsocFw3 = yes ProbSocModMw3 = none SigDoseProbsocMw3 = no MainEffPrbsocMw3 = age radhlw3 shjobw3 hmcareMedFw3 = age illw3 hmcareMedMw3 = age illw3 hmcareModFw3 = none SigdoseHmcareFw3 = no hmcareModMw3 = none MainEffhmcareMw3 = none SigDoseHmcareMw3 = no wkMedMw3 = bf8 age illw3 ageXillw3 wkModFw3 = none wkModMw3 = none MainEffwkFw3 = age MainEffwkMw3 = workM: age bf8 illw3 shjobw3 SigDoseWKMw3 = no SigDoseWkFw3 = no 592 . 593 . 594 . * X * missing the number of main effects in the trimmed models 595 . 596 . //////////////////////////////////////////////////////////////////////// > *--------- Chunk 4 Dose social problem impact relationship HP2probsoc 597 . 598 . 599 . title "3. H1 part 2 wave 1 Dose - HP2probsoc impact tested" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 3. H1 part 2 wave 1 Dose - HP2probsoc impact tested ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:14:16 ***** ******************************************************************************* ******************************************************************************* 600 . 601 . forvalues j=1/1 { 2. set more off 3. 602 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 603 . foreach var in HP2probsoc { 5. forvalues k=1/2 { 6. di _skip(4) 7. di as input "Full main model for `var' for wave= `j' " 8. di _skip(4) 9. di as input "chunk 4 H1 test:Gender= `k' model Wave = `j' for `e(depvar > )' " 10. di _skip(4) 11. 604 . 605 . xi: logistic `var' age i.educ occ1w`j'-occ8w`j' /// > marrw`j'1- marrw`j'3 marrw`j'5-marrw`j'6 inc1w`j'-inc4w`j' // > / > radhlw`j' havmil avgcumdosew`j' `w`j'bf' /// > deaw`j' dvcew`j' sepaw`j' accdw`j' movew`j' /// > illw`j' shfamw`j' shhlw`j' shjobw`j' shrelaw`j' suprtw`j' suc > hrw`j' /// > havmilsq if gender==`k', coef difficult iterate(50) 12. estat class 13. estat gof 14. fitstat 15. } 16. } 17. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2probsoc for wave= 1 chunk 4 H1 test:Gender= 1 model Wave = 1 for hp2hmcare i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: _Ieduc_4 != 0 predicts failure perfectly _Ieduc_4 dropped and 13 obs not used note: _Ieduc_7 != 0 predicts failure perfectly _Ieduc_7 dropped and 4 obs not used note: _Ieduc_8 != 0 predicts failure perfectly _Ieduc_8 dropped and 2 obs not used note: occ6w1 != 0 predicts failure perfectly occ6w1 dropped and 5 obs not used note: occ7w1 != 0 predicts failure perfectly occ7w1 dropped and 4 obs not used note: marrw12 != 0 predicts failure perfectly marrw12 dropped and 4 obs not used note: bf15m != 0 predicts failure perfectly bf15m dropped and 19 obs not used note: _Ieduc_6 omitted because of collinearity note: marrw16 omitted because of collinearity Logistic regression Number of obs = 288 LR chi2(43) = 122.03 Prob > chi2 = 0.0000 Log likelihood = -56.842934 Pseudo R2 = 0.5177 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0932157 .0488397 1.91 0.056 -.0025083 .1889398 _Ieduc_2 | -1.534787 1.373444 -1.12 0.264 -4.226687 1.157114 _Ieduc_3 | -2.172657 .8355094 -2.60 0.009 -3.810226 -.5350887 _Ieduc_4 | 0 (omitted) _Ieduc_5 | -1.97152 1.083493 -1.82 0.069 -4.095126 .1520862 _Ieduc_6 | 0 (omitted) _Ieduc_7 | 0 (omitted) _Ieduc_8 | 0 (omitted) occ1w1 | .4116733 2.029299 0.20 0.839 -3.565679 4.389026 occ2w1 | 1.622509 2.037766 0.80 0.426 -2.371439 5.616456 occ3w1 | 1.707699 2.267987 0.75 0.451 -2.737474 6.152872 occ4w1 | .8391363 2.160815 0.39 0.698 -3.395982 5.074255 occ5w1 | .2067084 2.315894 0.09 0.929 -4.33236 4.745777 occ6w1 | 0 (omitted) occ7w1 | 0 (omitted) occ8w1 | 1.657954 2.464035 0.67 0.501 -3.171466 6.487374 marrw11 | 8.254014 1282.599 0.01 0.995 -2505.595 2522.103 marrw12 | 0 (omitted) marrw13 | 6.455541 1282.6 0.01 0.996 -2507.394 2520.305 marrw15 | 45.60071 8350.823 0.01 0.996 -16321.71 16412.91 marrw16 | 0 (omitted) inc1w1 | -4.624348 2.584817 -1.79 0.074 -9.690496 .4418005 inc2w1 | -1.315065 2.130506 -0.62 0.537 -5.490781 2.860651 inc3w1 | -.8805899 2.024434 -0.43 0.664 -4.848408 3.087229 inc4w1 | -3.930393 2.652434 -1.48 0.138 -9.129068 1.268283 radhlw1 | -.003808 .0098127 -0.39 0.698 -.0230405 .0154246 havmil | .0037068 .0075299 0.49 0.623 -.0110516 .0184652 avgcumdosew1 | .1390885 .1022549 1.36 0.174 -.0613275 .3395044 bf1 | .0837527 .0660851 1.27 0.205 -.0457717 .2132771 bf4 | -.1765285 .2854179 -0.62 0.536 -.7359372 .3828802 bf9 | .0121714 .045033 0.27 0.787 -.0760915 .1004344 bf10 | -.0731627 .0326403 -2.24 0.025 -.1371365 -.0091889 bf11 | -.0133658 .0915598 -0.15 0.884 -.1928197 .1660881 bf4m | -.1819261 .2441274 -0.75 0.456 -.6604069 .2965547 bf15m | 0 (omitted) bf20 | -.0741254 .0559936 -1.32 0.186 -.1838707 .03562 bf22 | .0002567 .0001585 1.62 0.105 -.000054 .0005673 bf30 | -.0003312 .0005407 -0.61 0.540 -.0013911 .0007286 bf40 | -.1589223 .268107 -0.59 0.553 -.6844023 .3665577 deaw1 | .073587 .5137846 0.14 0.886 -.9334124 1.080586 dvcew1 | -33.19213 8251.737 -0.00 0.997 -16206.3 16139.92 sepaw1 | 21.75582 8052.548 0.00 0.998 -15760.95 15804.46 accdw1 | 1.124538 1.440947 0.78 0.435 -1.699667 3.948743 movew1 | .4856006 .9962789 0.49 0.626 -1.46707 2.438271 illw1 | .1109162 .620821 0.18 0.858 -1.105871 1.327703 shfamw1 | -.0009812 .0110299 -0.09 0.929 -.0225994 .020637 shhlw1 | .0236901 .0186204 1.27 0.203 -.0128052 .0601853 shjobw1 | .0162037 .0119225 1.36 0.174 -.0071639 .0395713 shrelaw1 | -.0240964 .0119609 -2.01 0.044 -.0475394 -.0006534 suprtw1 | .0381166 .0331693 1.15 0.250 -.026894 .1031273 suchrw1 | -.0419264 .049205 -0.85 0.394 -.1383664 .0545135 havmilsq | -5.12e-06 .0000107 -0.48 0.632 -.0000261 .0000159 _cons | -6.454218 1282.606 -0.01 0.996 -2520.316 2507.407 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 26 6 | 32 - | 15 241 | 256 -----------+--------------------------+----------- Total | 41 247 | 288 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 63.41% Specificity Pr( -|~D) 97.57% Positive predictive value Pr( D| +) 81.25% Negative predictive value Pr(~D| -) 94.14% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.43% False - rate for true D Pr( -| D) 36.59% False + rate for classified + Pr(~D| +) 18.75% False - rate for classified - Pr( D| -) 5.86% -------------------------------------------------- Correctly classified 92.71% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 288 number of covariate patterns = 288 Pearson chi2(244) = 239.99 Prob > chi2 = 0.5605 Measures of Fit for logistic of HP2probsoc Log-Lik Intercept Only: -117.857 Log-Lik Full Model: -56.843 D(235): 113.686 LR(43): 122.029 Prob > LR: 0.000 McFadden's R2: 0.518 McFadden's Adj R2: 0.068 Maximum Likelihood R2: 0.345 Cragg & Uhler's R2: 0.618 McKelvey and Zavoina's R2: 0.804 Efron's R2: 0.535 Variance of y*: 16.818 Variance of error: 3.290 Count R2: 0.927 Adj Count R2: 0.488 AIC: 0.763 AIC*n: 219.686 BIC: -1217.110 BIC': 121.479 Full main model for HP2probsoc for wave= 1 chunk 4 H1 test:Gender= 2 model Wave = 1 for HP2probsoc i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: bf15m != 0 predicts failure perfectly bf15m dropped and 11 obs not used note: _Ieduc_8 omitted because of collinearity convergence not achieved Logistic regression Number of obs = 347 LR chi2(49) = 174.17 Prob > chi2 = 0.0000 Log likelihood = -92.74522 Pseudo R2 = 0.4843 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1339222 .0304099 4.40 0.000 .07432 .1935245 _Ieduc_2 | -14.8519 3.844577 -3.86 0.000 -22.38714 -7.316671 _Ieduc_3 | -15.07441 3.702435 -4.07 0.000 -22.33105 -7.817769 _Ieduc_4 | -15.0354 3.71929 -4.04 0.000 -22.32507 -7.745721 _Ieduc_5 | -15.31894 3.642287 -4.21 0.000 -22.45769 -8.180187 _Ieduc_6 | -15.97033 3.721806 -4.29 0.000 -23.26493 -8.675722 _Ieduc_7 | -14.57197 4.858518 -3.00 0.003 -24.09449 -5.049449 _Ieduc_8 | 0 (omitted) occ1w1 | -2.855851 2.862357 -1.00 0.318 -8.465967 2.754266 occ2w1 | -2.736414 2.877102 -0.95 0.342 -8.375431 2.902603 occ3w1 | -3.494041 2.901139 -1.20 0.228 -9.180168 2.192087 occ4w1 | -3.424951 2.969387 -1.15 0.249 -9.244843 2.394942 occ5w1 | -4.951372 3.138113 -1.58 0.115 -11.10196 1.199216 occ6w1 | -4.529022 3.152748 -1.44 0.151 -10.70829 1.650251 occ7w1 | -2.040185 2.865556 -0.71 0.476 -7.656572 3.576202 occ8w1 | -1.038951 3.012173 -0.34 0.730 -6.942703 4.8648 marrw11 | -3.017556 2.506541 -1.20 0.229 -7.930285 1.895174 marrw12 | -1.448781 2.689444 -0.54 0.590 -6.719994 3.822432 marrw13 | -2.818836 2.317521 -1.22 0.224 -7.361093 1.723421 marrw15 | -4.022412 2.774385 -1.45 0.147 -9.460106 1.415282 marrw16 | -3.024951 2.89161 -1.05 0.296 -8.692402 2.6425 inc1w1 | 1.652153 2.826464 0.58 0.559 -3.887614 7.191921 inc2w1 | .7523029 2.773835 0.27 0.786 -4.684314 6.188919 inc3w1 | .8186418 2.79302 0.29 0.769 -4.655577 6.29286 inc4w1 | 1.416767 2.932929 0.48 0.629 -4.331668 7.165203 radhlw1 | .0078893 .0071763 1.10 0.272 -.0061759 .0219546 havmil | -.0016014 .0088347 -0.18 0.856 -.018917 .0157143 avgcumdosew1 | .9064706 .4019796 2.26 0.024 .1186049 1.694336 bf1 | .035394 .0349482 1.01 0.311 -.0331032 .1038912 bf4 | -.3079106 .2679557 -1.15 0.251 -.8330942 .217273 bf9 | -.0727705 .032505 -2.24 0.025 -.1364792 -.0090617 bf10 | -.0269165 .0168597 -1.60 0.110 -.059961 .0061279 bf11 | -.0076362 .0541236 -0.14 0.888 -.1137164 .0984441 bf4m | .0576549 .2545602 0.23 0.821 -.441274 .5565838 bf15m | 0 (omitted) bf20 | -.0388067 .0274447 -1.41 0.157 -.0925973 .0149838 bf22 | .0001098 .0001019 1.08 0.281 -.0000899 .0003094 bf30 | .0003325 .0003748 0.89 0.375 -.000402 .001067 bf40 | -.1036356 .1492079 -0.69 0.487 -.3960777 .1888066 deaw1 | -.1970026 .3234174 -0.61 0.542 -.830889 .4368837 dvcew1 | 1.249232 2.862648 0.44 0.663 -4.361455 6.859919 sepaw1 | -.7111444 3.235109 -0.22 0.826 -7.051842 5.629553 accdw1 | -.7069674 1.124581 -0.63 0.530 -2.911106 1.497171 movew1 | .3733727 .7129786 0.52 0.601 -1.02404 1.770785 illw1 | -.2000586 .4406407 -0.45 0.650 -1.063699 .6635813 shfamw1 | -.014563 .009194 -1.58 0.113 -.0325829 .003457 shhlw1 | -.0149441 .0130685 -1.14 0.253 -.0405578 .0106697 shjobw1 | .0015703 .0091986 0.17 0.864 -.0164586 .0195992 shrelaw1 | -.0082661 .0085598 -0.97 0.334 -.0250429 .0085107 suprtw1 | .0056689 .0131922 0.43 0.667 -.0201873 .0315251 suchrw1 | -.0127124 .0182568 -0.70 0.486 -.0484951 .0230702 havmilsq | -1.64e-06 .0000208 -0.08 0.937 -.0000424 .0000391 _cons | 15.72467 . . . . . ------------------------------------------------------------------------------ Note: 1 failure and 1 success completely determined. Warning: convergence not achieved Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 46 12 | 58 - | 28 261 | 289 -----------+--------------------------+----------- Total | 74 273 | 347 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 62.16% Specificity Pr( -|~D) 95.60% Positive predictive value Pr( D| +) 79.31% Negative predictive value Pr(~D| -) 90.31% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 4.40% False - rate for true D Pr( -| D) 37.84% False + rate for classified + Pr(~D| +) 20.69% False - rate for classified - Pr( D| -) 9.69% -------------------------------------------------- Correctly classified 88.47% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 347 number of covariate patterns = 347 Pearson chi2(296) = 277.13 Prob > chi2 = 0.7779 Measures of Fit for logistic of HP2probsoc Log-Lik Intercept Only: -179.829 Log-Lik Full Model: -92.745 D(294): 185.490 LR(49): 174.168 Prob > LR: 0.000 McFadden's R2: 0.484 McFadden's Adj R2: 0.190 Maximum Likelihood R2: 0.395 Cragg & Uhler's R2: 0.612 McKelvey and Zavoina's R2: 0.757 Efron's R2: 0.511 Variance of y*: 13.536 Variance of error: 3.290 Count R2: 0.885 Adj Count R2: 0.459 AIC: 0.840 AIC*n: 291.490 BIC: -1534.211 BIC': 112.449 606 . 607 . *-----Chunk 4 dose3 social problem impact--------------no sig dose main effec > t-- 608 . title4 "Male trimmed models of dose and HP2probsoc relationship in wave 1" ------------------------------------------------------------------------------- Male trimmed models of dose and HP2probsoc relationship in wave 1 ------------------------------------------------------------------------------- 609 . * male models 610 . forvalues j=1/1 { 2. set more off 3. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 4. title4 "trimmed HP2probsoc main effects models wave 1 for H1 part 2 with d > ose ns" 5. title4 "wave 1 dose HP2probsoc relationship but avgcumdosew`j': Dose not s > ignif" 6. sw, pr(.1): logit HP2probsoc age radhlw1 accdw`j' `w`j'bf' shjobw`j' hav > milsq /// > avgcumdosew`j' shhlw`j' shrelaw`j' if gender==1 7. estat class 8. estat gof 9. fitstat 10. } ------------------------------------------------------------------------------- trimmed HP2probsoc main effects models wave 1 for H1 part 2 with dose ns ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- wave 1 dose HP2probsoc relationship but avgcumdosew1: Dose not signif ------------------------------------------------------------------------------- note: bf15m dropped because of estimability note: o.bf15m dropped because of estimability note: 21 obs. dropped because of estimability begin with full model p = 0.8569 >= 0.1000 removing accdw1 p = 0.8345 >= 0.1000 removing bf4 p = 0.6090 >= 0.1000 removing havmilsq p = 0.5895 >= 0.1000 removing radhlw1 p = 0.6078 >= 0.1000 removing bf40 p = 0.5292 >= 0.1000 removing bf20 p = 0.7236 >= 0.1000 removing bf1 p = 0.3440 >= 0.1000 removing bf11 p = 0.3090 >= 0.1000 removing bf10 p = 0.2900 >= 0.1000 removing bf9 p = 0.3215 >= 0.1000 removing bf30 p = 0.2860 >= 0.1000 removing bf22 p = 0.1710 >= 0.1000 removing avgcumdosew1 p = 0.1317 >= 0.1000 removing shhlw1 p = 0.1690 >= 0.1000 removing shrelaw1 p = 0.1067 >= 0.1000 removing shjobw1 Logistic regression Number of obs = 319 LR chi2(2) = 71.57 Prob > chi2 = 0.0000 Log likelihood = -86.576328 Pseudo R2 = 0.2925 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0381014 .0181494 2.10 0.036 .0025293 .0736736 bf4m | -.2064677 .0337163 -6.12 0.000 -.2725504 -.140385 _cons | -.2900951 1.272585 -0.23 0.820 -2.784316 2.204126 ------------------------------------------------------------------------------ Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 13 6 | 19 - | 28 272 | 300 -----------+--------------------------+----------- Total | 41 278 | 319 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 31.71% Specificity Pr( -|~D) 97.84% Positive predictive value Pr( D| +) 68.42% Negative predictive value Pr(~D| -) 90.67% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.16% False - rate for true D Pr( -| D) 68.29% False + rate for classified + Pr(~D| +) 31.58% False - rate for classified - Pr( D| -) 9.33% -------------------------------------------------- Correctly classified 89.34% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 319 number of covariate patterns = 220 Pearson chi2(217) = 200.09 Prob > chi2 = 0.7885 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -122.361 Log-Lik Full Model: -86.576 D(316): 173.153 LR(2): 71.569 Prob > LR: 0.000 McFadden's R2: 0.292 McFadden's Adj R2: 0.268 Maximum Likelihood R2: 0.201 Cragg & Uhler's R2: 0.375 McKelvey and Zavoina's R2: 0.372 Efron's R2: 0.293 Variance of y*: 5.240 Variance of error: 3.290 Count R2: 0.893 Adj Count R2: 0.171 AIC: 0.562 AIC*n: 179.153 BIC: -1648.648 BIC': -60.039 611 . title4 "Trimmed male main effects model" ------------------------------------------------------------------------------- Trimmed male main effects model ------------------------------------------------------------------------------- 612 . forvalues j=1/1 { 2. title "trimmed HP2probsoc main effects models wave `j'" " for H1 part 2 wi > th dose ns" 3. title2 "Wave `j dose HP2work relationship but avgcumdosew`j': Dose not sig > nif" 4. } ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** trimmed HP2probsoc main effects models wave 1 ***** ***** for H1 part 2 with dose ns ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:14:47 ***** ******************************************************************************* ******************************************************************************* ------------------------------------------------------------------------------- title2: Wave `j dose HP2work relationship but avgcumdosew1: Dose not signif Date and time: 18 Jun 2012 18:14:47 Working directory: /Users/robertyaffee/ > Documents/data/research/chwk/phase3/Htests/h1tests/h1pt2 Stata data file: chwide16june2012.dta > has 2382 variables and 703 observations Wave `j dose HP2work relationship but avgcumdosew1: Dose not signif ------------------------------------------------------------------------------- 613 . 614 . logit HP2probsoc age radhlw1 bf4m if gender==1 Iteration 0: log likelihood = -125.15243 Iteration 1: log likelihood = -98.286147 Iteration 2: log likelihood = -88.284687 Iteration 3: log likelihood = -87.790699 Iteration 4: log likelihood = -87.789842 Iteration 5: log likelihood = -87.789842 Logistic regression Number of obs = 340 LR chi2(3) = 74.73 Prob > chi2 = 0.0000 Log likelihood = -87.789842 Pseudo R2 = 0.2985 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0408433 .0180763 2.26 0.024 .0054145 .0762722 radhlw1 | .0025885 .0057387 0.45 0.652 -.0086592 .0138361 bf4m | -.1982136 .0358985 -5.52 0.000 -.2685733 -.1278538 _cons | -.7777696 1.352662 -0.57 0.565 -3.428938 1.873399 ------------------------------------------------------------------------------ 615 . 616 . 617 . scalar SigdoseMw1 = "no" 618 . scalar MainEffPrbsocMw1 = "age bf4m" 619 . 620 . 621 . foreach var in bf4 age radhlw1 shjobw1 shrelaw1 { 2. cap gen `var'Xd1= `var'*avgcumdosew1 3. } 622 . 623 . 624 . 625 . forvalues j=1/1 { 2. set more off 3. title "Main effects Dose ProbSoc model for males" 4. logit HP2probsoc age avgcumdosew1 radhlw1 shjobw`j' /// > ageXd1 radhlw1Xd1 shjobw1Xd1 shrelaw1Xd1 if gender==1 5. estat class 6. estat gof 7. fitstat 8. } ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Main effects Dose ProbSoc model for males ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:14:48 ***** ******************************************************************************* ******************************************************************************* Iteration 0: log likelihood = -125.15243 Iteration 1: log likelihood = -96.957486 Iteration 2: log likelihood = -91.634 Iteration 3: log likelihood = -90.570723 Iteration 4: log likelihood = -90.307985 Iteration 5: log likelihood = -90.296231 Iteration 6: log likelihood = -90.296206 Iteration 7: log likelihood = -90.296206 Logistic regression Number of obs = 340 LR chi2(8) = 69.71 Prob > chi2 = 0.0000 Log likelihood = -90.296206 Pseudo R2 = 0.2785 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0348678 .0289544 1.20 0.229 -.0218818 .0916174 avgcumdosew1 | -13.34838 6.796946 -1.96 0.050 -26.67015 -.0266126 radhlw1 | .0078565 .0063701 1.23 0.217 -.0046287 .0203416 shjobw1 | .0100379 .010521 0.95 0.340 -.0105828 .0306587 ageXd1 | .1331427 .1074383 1.24 0.215 -.0774326 .3437179 radhlw1Xd1 | .0060028 .0144906 0.41 0.679 -.0223983 .0344039 shjobw1Xd1 | .0795578 .0390013 2.04 0.041 .0031166 .155999 shrelaw1Xd1 | -.0381989 .015358 -2.49 0.013 -.0683 -.0080977 _cons | -4.518141 1.78078 -2.54 0.011 -8.008406 -1.027877 ------------------------------------------------------------------------------ Note: 2 failures and 0 successes completely determined. Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 10 3 | 13 - | 31 296 | 327 -----------+--------------------------+----------- Total | 41 299 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 24.39% Specificity Pr( -|~D) 99.00% Positive predictive value Pr( D| +) 76.92% Negative predictive value Pr(~D| -) 90.52% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.00% False - rate for true D Pr( -| D) 75.61% False + rate for classified + Pr(~D| +) 23.08% False - rate for classified - Pr( D| -) 9.48% -------------------------------------------------- Correctly classified 90.00% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 340 number of covariate patterns = 331 Pearson chi2(322) = 260.41 Prob > chi2 = 0.9950 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -125.152 Log-Lik Full Model: -90.296 D(331): 180.592 LR(8): 69.712 Prob > LR: 0.000 McFadden's R2: 0.279 McFadden's Adj R2: 0.207 Maximum Likelihood R2: 0.185 Cragg & Uhler's R2: 0.356 McKelvey and Zavoina's R2: 0.588 Efron's R2: 0.248 Variance of y*: 7.976 Variance of error: 3.290 Count R2: 0.900 Adj Count R2: 0.171 AIC: 0.584 AIC*n: 198.592 BIC: -1748.789 BIC': -23.081 626 . 627 . title4 "male wave 1 moderators of dose and hmcare" ------------------------------------------------------------------------------- male wave 1 moderators of dose and hmcare ------------------------------------------------------------------------------- 628 . scalar PrbsocModMw1 = "shjobw1Xd1 shrelaw1Xd1" 629 . 630 . forvalues j=1/1 { 2. set more off 3. logit HP2probsoc age radhlw1 shjobw`j' shrelaw1 /// > avgcumdosew`j' shrelaw`j'Xd1 /// > shjobw1Xd1 if gender==1 4. estat class 5. estat gof 6. fitstat 7. } Iteration 0: log likelihood = -125.15243 Iteration 1: log likelihood = -98.199364 Iteration 2: log likelihood = -91.871369 Iteration 3: log likelihood = -91.382931 Iteration 4: log likelihood = -91.279919 Iteration 5: log likelihood = -91.277661 Iteration 6: log likelihood = -91.277658 Logistic regression Number of obs = 340 LR chi2(7) = 67.75 Prob > chi2 = 0.0000 Log likelihood = -91.277658 Pseudo R2 = 0.2707 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .065245 .0168011 3.88 0.000 .0323155 .0981744 radhlw1 | .0097595 .0055022 1.77 0.076 -.0010247 .0205437 shjobw1 | .0098453 .0120807 0.81 0.415 -.0138324 .033523 shrelaw1 | .0002189 .0069348 0.03 0.975 -.0133731 .0138109 avgcumdosew1 | -5.108591 3.138225 -1.63 0.104 -11.2594 1.042216 shrelaw1Xd1 | -.0269782 .0164348 -1.64 0.101 -.0591898 .0052334 shjobw1Xd1 | .0726131 .0410722 1.77 0.077 -.0078869 .1531132 _cons | -6.378223 1.306394 -4.88 0.000 -8.938707 -3.817739 ------------------------------------------------------------------------------ Note: 2 failures and 0 successes completely determined. Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 9 4 | 13 - | 32 295 | 327 -----------+--------------------------+----------- Total | 41 299 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 21.95% Specificity Pr( -|~D) 98.66% Positive predictive value Pr( D| +) 69.23% Negative predictive value Pr(~D| -) 90.21% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.34% False - rate for true D Pr( -| D) 78.05% False + rate for classified + Pr(~D| +) 30.77% False - rate for classified - Pr( D| -) 9.79% -------------------------------------------------- Correctly classified 89.41% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 340 number of covariate patterns = 331 Pearson chi2(323) = 280.63 Prob > chi2 = 0.9572 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -125.152 Log-Lik Full Model: -91.278 D(332): 182.555 LR(7): 67.750 Prob > LR: 0.000 McFadden's R2: 0.271 McFadden's Adj R2: 0.207 Maximum Likelihood R2: 0.181 Cragg & Uhler's R2: 0.347 McKelvey and Zavoina's R2: 0.590 Efron's R2: 0.242 Variance of y*: 8.026 Variance of error: 3.290 Count R2: 0.894 Adj Count R2: 0.122 AIC: 0.584 AIC*n: 198.555 BIC: -1752.655 BIC': -26.947 631 . scalar SigDoseProbsocMw1 = "no" 632 . * xx no signific radhwl3 by dose effect 633 . * xx for males no signif dose social problem effect 634 . * xx for males no significant moderator in dose social problem effect 635 . 636 . * female models 637 . 638 . 639 . 640 . 641 . 642 . title4 "H1 pt2 wave 1 trimmed female moderator model with basis functions" ------------------------------------------------------------------------------- H1 pt2 wave 1 trimmed female moderator model with basis functions ------------------------------------------------------------------------------- 643 . forvalues j=1/1 { 2. set more off 3. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 4. title "trimmed HP2probsoc main effects models wave 1 for H1 part 2 " "Dose > is signif Females" 5. title "wave 1 dose HP2probsoc relationship but avgcumdosew`j': Dose signif > " 6. sw, pr(.1): logit HP2probsoc age radhlw1 illw`j' `w1bf' //// > shrelaw`j' avgcumdosew`j' if gender==2 7. estat class 8. estat gof 9. fitstat 10. } ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** trimmed HP2probsoc main effects models wave 1 for H1 part 2 ***** ***** Dose is signif Females ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:14:51 ***** ******************************************************************************* ******************************************************************************* ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** wave 1 dose HP2probsoc relationship but avgcumdosew1: Dose signif ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:14:51 ***** ******************************************************************************* ******************************************************************************* note: bf15m dropped because of estimability note: o.bf15m dropped because of estimability note: 11 obs. dropped because of estimability begin with full model p = 0.5688 >= 0.1000 removing illw1 p = 0.4993 >= 0.1000 removing bf40 p = 0.5410 >= 0.1000 removing bf22 p = 0.4121 >= 0.1000 removing bf11 p = 0.4390 >= 0.1000 removing bf10 p = 0.2876 >= 0.1000 removing bf1 p = 0.4202 >= 0.1000 removing bf20 p = 0.3289 >= 0.1000 removing bf30 p = 0.2425 >= 0.1000 removing bf9 p = 0.1125 >= 0.1000 removing bf4m Logistic regression Number of obs = 351 LR chi2(5) = 134.88 Prob > chi2 = 0.0000 Log likelihood = -113.34071 Pseudo R2 = 0.3731 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0966615 .0183518 5.27 0.000 .0606925 .1326304 radhlw1 | .0116609 .004855 2.40 0.016 .0021453 .0211765 avgcumdosew1 | .6600342 .2959902 2.23 0.026 .0799041 1.240164 shrelaw1 | -.0108373 .005085 -2.13 0.033 -.0208036 -.0008709 bf4 | -.1838833 .0354969 -5.18 0.000 -.2534559 -.1143107 _cons | -5.736784 1.146357 -5.00 0.000 -7.983602 -3.489965 ------------------------------------------------------------------------------ Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 41 14 | 55 - | 33 263 | 296 -----------+--------------------------+----------- Total | 74 277 | 351 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 55.41% Specificity Pr( -|~D) 94.95% Positive predictive value Pr( D| +) 74.55% Negative predictive value Pr(~D| -) 88.85% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.05% False - rate for true D Pr( -| D) 44.59% False + rate for classified + Pr(~D| +) 25.45% False - rate for classified - Pr( D| -) 11.15% -------------------------------------------------- Correctly classified 86.61% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 351 number of covariate patterns = 350 Pearson chi2(344) = 376.81 Prob > chi2 = 0.1078 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -180.782 Log-Lik Full Model: -113.341 D(345): 226.681 LR(5): 134.883 Prob > LR: 0.000 McFadden's R2: 0.373 McFadden's Adj R2: 0.340 Maximum Likelihood R2: 0.319 Cragg & Uhler's R2: 0.496 McKelvey and Zavoina's R2: 0.548 Efron's R2: 0.411 Variance of y*: 7.277 Variance of error: 3.290 Count R2: 0.866 Adj Count R2: 0.365 AIC: 0.680 AIC*n: 238.681 BIC: -1795.290 BIC': -105.579 644 . 645 . 646 . scalar SigDoseProbsocFw1 = "yes" 647 . scalar MainEffProbSocFw1 = "age radhlw1 avgcumdosew1 shrelaw1 bf4" 648 . 649 . foreach var in b4 shrelaw1 { 2. cap gen `var'Xd1 = `var'*avgcumdosew1 3. } 650 . 651 . * testing the female moderator model with basis functions 652 . forvalues j=1/1 { 2. set more off 3. local w1bf bf1 bf4 bf9 bf10 bf11 bf14 bf15m bf20 bf22 bf30 bf40 4. title "trimmed HP2socprob main effects wv 1 for Hyp1 pt 2" "dose is signi > f for females" 5. title "wave 1 dose HP2socprob relationship but avgcumdosew`j'" " Dose is s > ignif for females" 6. sw, pr(.1): logit HP2probsoc age radhlw1 illw`j' `w1bf' //// > shrelaw`j' avgcumdosew`j' bf4Xd1 shrelaw1Xd1 /// > if gender==2 7. estat class 8. estat gof 9. fitstat 10. } ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** trimmed HP2socprob main effects wv 1 for Hyp1 pt 2 ***** ***** dose is signif for females ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:15:08 ***** ******************************************************************************* ******************************************************************************* ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** wave 1 dose HP2socprob relationship but avgcumdosew1 ***** ***** Dose is signif for females ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:15:08 ***** ******************************************************************************* ******************************************************************************* note: bf15m dropped because of estimability note: o.bf15m dropped because of estimability note: 11 obs. dropped because of estimability begin with full model p = 0.8198 >= 0.1000 removing shrelaw1Xd1 p = 0.4967 >= 0.1000 removing illw1 p = 0.4970 >= 0.1000 removing bf14 p = 0.4352 >= 0.1000 removing bf1 p = 0.4028 >= 0.1000 removing bf11 p = 0.3720 >= 0.1000 removing bf10 p = 0.3228 >= 0.1000 removing bf20 p = 0.3123 >= 0.1000 removing bf30 p = 0.2761 >= 0.1000 removing bf9 p = 0.1462 >= 0.1000 removing bf4Xd1 p = 0.1528 >= 0.1000 removing bf40 p = 0.3685 >= 0.1000 removing bf22 Logistic regression Number of obs = 351 LR chi2(5) = 134.88 Prob > chi2 = 0.0000 Log likelihood = -113.34071 Pseudo R2 = 0.3731 ------------------------------------------------------------------------------ HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0966615 .0183518 5.27 0.000 .0606925 .1326304 radhlw1 | .0116609 .004855 2.40 0.016 .0021453 .0211765 avgcumdosew1 | .6600342 .2959902 2.23 0.026 .0799041 1.240164 shrelaw1 | -.0108373 .005085 -2.13 0.033 -.0208036 -.0008709 bf4 | -.1838833 .0354969 -5.18 0.000 -.2534559 -.1143107 _cons | -5.736784 1.146357 -5.00 0.000 -7.983602 -3.489965 ------------------------------------------------------------------------------ Logistic model for HP2probsoc -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 41 14 | 55 - | 33 263 | 296 -----------+--------------------------+----------- Total | 74 277 | 351 Classified + if predicted Pr(D) >= .5 True D defined as HP2probsoc != 0 -------------------------------------------------- Sensitivity Pr( +| D) 55.41% Specificity Pr( -|~D) 94.95% Positive predictive value Pr( D| +) 74.55% Negative predictive value Pr(~D| -) 88.85% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.05% False - rate for true D Pr( -| D) 44.59% False + rate for classified + Pr(~D| +) 25.45% False - rate for classified - Pr( D| -) 11.15% -------------------------------------------------- Correctly classified 86.61% -------------------------------------------------- Logistic model for HP2probsoc, goodness-of-fit test number of observations = 351 number of covariate patterns = 350 Pearson chi2(344) = 376.81 Prob > chi2 = 0.1078 Measures of Fit for logit of HP2probsoc Log-Lik Intercept Only: -180.782 Log-Lik Full Model: -113.341 D(345): 226.681 LR(5): 134.883 Prob > LR: 0.000 McFadden's R2: 0.373 McFadden's Adj R2: 0.340 Maximum Likelihood R2: 0.319 Cragg & Uhler's R2: 0.496 McKelvey and Zavoina's R2: 0.548 Efron's R2: 0.411 Variance of y*: 7.277 Variance of error: 3.290 Count R2: 0.866 Adj Count R2: 0.365 AIC: 0.680 AIC*n: 238.681 BIC: -1795.290 BIC': -105.579 653 . 654 . scalar ProbsocModFw1 = "none" 655 . 656 . 657 . title4 "H1 pt 2 wave 1 testing for mediators for males" ------------------------------------------------------------------------------- H1 pt 2 wave 1 testing for mediators for males ------------------------------------------------------------------------------- 658 . * Male mediator dose social problem response models 659 . 660 . 661 . 662 . 663 . glm age avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 664 . glm HP2probsoc age if gender==1, fam(bin) irls link(probit) scale(dev) Iteration 1: deviance = 230.0069 Iteration 2: deviance = 223.5454 Iteration 3: deviance = 223.2469 Iteration 4: deviance = 223.2456 Iteration 5: deviance = 223.2456 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 223.2456447 (1/df) Deviance = .6604901 Pearson = 331.0340472 (1/df) Pearson = .9793907 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1746.938 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0392343 .0064286 6.10 0.000 .0266344 .0518342 _cons | -3.234772 .3587306 -9.02 0.000 -3.937871 -2.531673 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 665 . 666 . glm radhlw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1710.3417 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1378.645 Deviance = 465981.8893 (1/df) Deviance = 1378.645 Pearson = 465981.8893 (1/df) Pearson = 1378.645 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.0726 Log likelihood = -1710.341694 BIC = 464011.7 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 2.398285 1.208412 1.98 0.047 .0298407 4.766729 _cons | 44.66477 2.079384 21.48 0.000 40.58925 48.74029 ------------------------------------------------------------------------------ 667 . glm HP2probsoc radhlw1 if gender==1,fam(bin) irls link(probit) scale(dev) Iteration 1: deviance = 234.6563 Iteration 2: deviance = 229.8984 Iteration 3: deviance = 229.7634 Iteration 4: deviance = 229.7633 Iteration 5: deviance = 229.7633 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 229.7632837 (1/df) Deviance = .679773 Pearson = 344.9083225 (1/df) Pearson = 1.020439 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1740.42 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0108869 .0020577 5.29 0.000 .0068539 .0149198 _cons | -1.765962 .1445027 -12.22 0.000 -2.049182 -1.482742 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 668 . 669 . glm shjobw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1710.378 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1378.939 Deviance = 466081.3297 (1/df) Deviance = 1378.939 Pearson = 466081.3297 (1/df) Pearson = 1378.939 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.07281 Log likelihood = -1710.377969 BIC = 464111.1 ------------------------------------------------------------------------------ | OIM shjobw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 1.461274 1.208541 1.21 0.227 -.907423 3.82997 _cons | 48.94341 2.079606 23.53 0.000 44.86746 53.01936 ------------------------------------------------------------------------------ 670 . glm HP2probsoc shjobw1 if gender==1,fam(bin) irls link(probit) scale(dev) Iteration 1: deviance = 227.0059 Iteration 2: deviance = 218.5961 Iteration 3: deviance = 218.011 Iteration 4: deviance = 218.0073 Iteration 5: deviance = 218.0072 Iteration 6: deviance = 218.0072 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 218.0072484 (1/df) Deviance = .6449919 Pearson = 380.0066656 (1/df) Pearson = 1.12428 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1752.176 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shjobw1 | .0149051 .0023429 6.36 0.000 .010313 .0194971 _cons | -2.086099 .1780182 -11.72 0.000 -2.435009 -1.73719 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 671 . 672 . glm shrelaw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1715.4166 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1420.421 Deviance = 480102.3567 (1/df) Deviance = 1420.421 Pearson = 480102.3567 (1/df) Pearson = 1420.421 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.10245 Log likelihood = -1715.416629 BIC = 478132.2 ------------------------------------------------------------------------------ | OIM shrelaw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.0649405 1.226584 -0.05 0.958 -2.469002 2.339121 _cons | 29.57199 2.110654 14.01 0.000 25.43518 33.7088 ------------------------------------------------------------------------------ 673 . glm HP2probsoc shrelaw1 if gender==1,fam(bin) irls link(probit) scale(dev) Iteration 1: deviance = 250.4419 Iteration 2: deviance = 249.8324 Iteration 3: deviance = 249.8314 Iteration 4: deviance = 249.8314 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 249.8314263 (1/df) Deviance = .7391462 Pearson = 340.1867245 (1/df) Pearson = 1.00647 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1720.352 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shrelaw1 | .0015698 .0019726 0.80 0.426 -.0022964 .0054361 _cons | -1.220468 .0980174 -12.45 0.000 -1.412578 -1.028357 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 674 . 675 . title4 "Shjobw1 can be a mediator with others" ------------------------------------------------------------------------------- Shjobw1 can be a mediator with others ------------------------------------------------------------------------------- 676 . glm radhlw1 shjobw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1679.4377 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 337 Scale parameter = 1152.894 Deviance = 388525.4251 (1/df) Deviance = 1152.894 Pearson = 388525.4251 (1/df) Pearson = 1152.894 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 9.896692 Log likelihood = -1679.437708 BIC = 386561.1 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shjobw1 | .4076599 .0497352 8.20 0.000 .3101807 .5051391 avgcumdosew1 | 1.802582 1.107442 1.63 0.104 -.3679637 3.973128 _cons | 24.71251 3.088883 8.00 0.000 18.65841 30.76661 ------------------------------------------------------------------------------ 677 . glm HP2probsoc shjobw1 avgcumdosew1 shjobw1Xd1 radhlw1 if gender==1,fam(bin) > /// > irls scale(dev) link(probit) Iteration 1: deviance = 219.4001 Iteration 2: deviance = 209.1683 Iteration 3: deviance = 208.2026 Iteration 4: deviance = 208.1798 Iteration 5: deviance = 208.1796 Iteration 6: deviance = 208.1796 Iteration 7: deviance = 208.1796 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 335 (IRLS EIM) Scale parameter = 1 Deviance = 208.179583 (1/df) Deviance = .6214316 Pearson = 348.7623267 (1/df) Pearson = 1.041082 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1744.517 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shjobw1 | .0101989 .0030577 3.34 0.001 .004206 .0161918 avgcumdosew1 | -.5046729 .500832 -1.01 0.314 -1.486286 .4769398 shjobw1Xd1 | .0071158 .0061813 1.15 0.250 -.0049993 .0192309 radhlw1 | .0070111 .0022234 3.15 0.002 .0026533 .0113688 _cons | -2.183344 .2368622 -9.22 0.000 -2.647586 -1.719103 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 678 . 679 . scalar ProbsocMedMw1 = "radhlw1" 680 . 681 . title4 "H1 pt2 wave 1 HP2probsoc female mediator tests" ------------------------------------------------------------------------------- H1 pt2 wave 1 HP2probsoc female mediator tests ------------------------------------------------------------------------------- 682 . 683 . // age is a possible female mediator 684 . glm age avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 685 . glm HP2probsoc age if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 289.3253 Iteration 2: deviance = 280.9528 Iteration 3: deviance = 280.5176 Iteration 4: deviance = 280.5162 Iteration 5: deviance = 280.5162 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 280.5161869 (1/df) Deviance = .7770531 Pearson = 406.2926804 (1/df) Pearson = 1.125464 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1847.363 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0683554 .007474 9.15 0.000 .0537067 .083004 _cons | -4.505136 .424587 -10.61 0.000 -5.337311 -3.672961 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 686 . 687 . // radhlw1 is a possible female mediator 688 . glm radhlw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1821.9477 Generalized linear models No. of obs = 362 Optimization : ML Residual df = 360 Scale parameter = 1385.301 Deviance = 498708.3025 (1/df) Deviance = 1385.301 Pearson = 498708.3025 (1/df) Pearson = 1385.301 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.07706 Log likelihood = -1821.947718 BIC = 496587.3 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.789972 3.562254 1.06 0.287 -3.191917 10.77186 _cons | 55.44948 2.293757 24.17 0.000 50.9538 59.94516 ------------------------------------------------------------------------------ 689 . glm HP2probsoc radhlw1 if gender==2,fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 347.1581 Iteration 2: deviance = 346.2367 Iteration 3: deviance = 346.2357 Iteration 4: deviance = 346.2357 Generalized linear models No. of obs = 362 Optimization : MQL Fisher scoring Residual df = 360 (IRLS EIM) Scale parameter = 1 Deviance = 346.2357277 (1/df) Deviance = .9617659 Pearson = 365.2532496 (1/df) Pearson = 1.014592 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1774.756 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0094598 .0021222 4.46 0.000 .0053004 .0136191 _cons | -1.413461 .1575297 -8.97 0.000 -1.722213 -1.104708 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 690 . 691 . // bf4 is a possible female mediator 692 . glm bf4 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 693 . glm HP2probsoc bf4 if gender==2,fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 295.3213 Iteration 2: deviance = 289.8751 Iteration 3: deviance = 289.6607 Iteration 4: deviance = 289.6591 Iteration 5: deviance = 289.6591 Iteration 6: deviance = 289.6591 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 289.6591069 (1/df) Deviance = .8023798 Pearson = 306.7680355 (1/df) Pearson = .849773 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1838.22 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1364145 .0148745 -9.17 0.000 -.165568 -.107261 _cons | .3956217 .1443369 2.74 0.006 .1127267 .6785167 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 694 . 695 . glm shrelaw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1801.2674 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 1202.293 Deviance = 434027.9437 (1/df) Deviance = 1202.293 Pearson = 434027.9437 (1/df) Pearson = 1202.293 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 9.935358 Log likelihood = -1801.267425 BIC = 431900.1 ------------------------------------------------------------------------------ | OIM shrelaw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 5.240218 3.317628 1.58 0.114 -1.262215 11.74265 _cons | 26.01592 2.133342 12.19 0.000 21.83464 30.19719 ------------------------------------------------------------------------------ 696 . glm HP2probsoc shrelaw1 if gender==2,fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 367.3685 Iteration 2: deviance = 367.117 Iteration 3: deviance = 367.117 Iteration 4: deviance = 367.117 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 367.1169775 (1/df) Deviance = 1.016945 Pearson = 362.9967567 (1/df) Pearson = 1.005531 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1760.762 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shrelaw1 | -.0003307 .0021782 -0.15 0.879 -.0045998 .0039385 _cons | -.8187957 .0962545 -8.51 0.000 -1.007451 -.6301402 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 697 . 698 . 699 . glm radhlw1 shjobw1 avgcumdosew1 shjobw1Xd1 if gender==2, fam(gaus) link(iden > tity) Iteration 0: log likelihood = -1821.765 Generalized linear models No. of obs = 362 Optimization : ML Residual df = 358 Scale parameter = 1391.635 Deviance = 498205.2021 (1/df) Deviance = 1391.635 Pearson = 498205.2021 (1/df) Pearson = 1391.635 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.0871 Log likelihood = -1821.765032 BIC = 496096 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shjobw1 | .0182877 .0685464 0.27 0.790 -.1160606 .1526361 avgcumdosew1 | 12.32024 14.83908 0.83 0.406 -16.76382 41.4043 shjobw1Xd1 | -.1018921 .1735559 -0.59 0.557 -.4420555 .2382713 _cons | 53.68631 4.758015 11.28 0.000 44.36077 63.01185 ------------------------------------------------------------------------------ 700 . glm HP2probsoc shjobw1 avgcumdosew1 shjobw1Xd1 if gender==2,fam(bin) irls sc > ale(dev) link(probit) Iteration 1: deviance = 341.3502 Iteration 2: deviance = 340.0594 Iteration 3: deviance = 340.039 Iteration 4: deviance = 340.0389 Iteration 5: deviance = 340.0389 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 359 (IRLS EIM) Scale parameter = 1 Deviance = 340.0389343 (1/df) Deviance = .9471837 Pearson = 349.6343388 (1/df) Pearson = .9739118 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1776.052 ------------------------------------------------------------------------------ | EIM HP2probsoc | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shjobw1 | .0078885 .0028001 2.82 0.005 .0024004 .0133767 avgcumdosew1 | 2.042963 .654931 3.12 0.002 .7593215 3.326604 shjobw1Xd1 | -.0171602 .00714 -2.40 0.016 -.0311544 -.0031661 _cons | -1.586799 .2154854 -7.36 0.000 -2.009142 -1.164455 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 701 . 702 . 703 . scalar SigDoseProbsocFw1 = "yes" 704 . scalar SigDoseProbsocMw1 = "no" 705 . 706 . 707 . scalar MainEffPrbsocMw1 = "age bf4m" 708 . scalar MainEffProbSocFw1 = "age radhlw1 avgcumdosew1 shrelaw1 bf4" 709 . 710 . scalar PrbsocModMw1 = "shjobw1Xd1 shrelaw1Xd1" 711 . scalar ProbsocModFw1 = "none" 712 . 713 . 714 . scalar ProbsocMedMw1 = "radhlw1" 715 . scalar ProbsocMedFw1 = "age bf4 " 716 . 717 . 718 . * male hp2spM w1 mediators: age 719 . * dose is not significant main effect for males 720 . * female hp2spF w1 mediators: age radhlw1 721 . * dose is not sig main effect for males 722 . 723 . 724 . 725 . 726 . 727 . title4 "3. Matrix summary for H1 pt2 wave 1 HP2probsoc Impact" ------------------------------------------------------------------------------- 3. Matrix summary for H1 pt2 wave 1 HP2probsoc Impact ------------------------------------------------------------------------------- 728 . matrix define spMw1 = J(1,8, 0) 729 . matrix define spFw1 = J(1,8, 0) 730 . matrix colnames spMw1= hypnum ptnum wave gender medsig numMAsig numModsig n > umMed 731 . matrix colnames spFw1= hypnum ptnum wave gender medsig numMAsig numModsig n > umMed 732 . 733 . matrix define spMw1= (1, 2, 1, 1, 1, 2, 2, 1 ) 734 . matrix define spFw1= (1, 2, 1, 2, 0, 5, 0, 2 ) 735 . matrix rowname spMw1 = spMw1 736 . matrix rowname spFw1 = spFw1 737 . matlist spMw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- spMw1 | 1 2 1 1 1 2 > 2 1 738 . matlist spFw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- spFw1 | 1 2 1 2 0 5 > 0 2 739 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1 \ hmcrMw1 \ hmcrFw1 \ sp > Mw1 \ spFw1 ) 740 . 741 . matlist H1pt2w1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 1 > 1 2 r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 2 > 0 2 spMw1 | 1 2 1 1 1 2 > 2 1 spFw1 | 1 2 1 2 0 5 > 0 2 742 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 743 . matrix rownames H1pt2w1 = wkMw1 wkFw1 hmcrMw1 hmcrFw1 socprbMw1 socprbF > w1 744 . matlist H1pt2w1 | hypnum ptnum wave gender medsig numMAsig > numModsig numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 hmcrMw1 | 1 2 1 1 0 2 > 0 1 hmcrFw1 | 1 2 1 2 0 2 > 0 2 socprbMw1 | 1 2 1 1 1 2 > 2 1 socprbFw1 | 1 2 1 2 0 5 > 0 2 745 . 746 . 747 . 748 . *xx significant dose effect for females 749 . 750 . *xx no female moderators for Dose Social problem impact relationship 751 . scalar list MainEffhmcrFw1 = age MainEffhmcrMw1 = age hmcrMedFw1 = age bf4 hmcrMedMw1 = radhlw1 MainEffwkFw1 = age MainEffwkMw1 = age MainEffVactnMw1 = age radhlw1 VactnMedFw1 = age illw1 radhlw1 VactnMedMw1 = age VacatnModFw1 = none MainEffVactnFw1 = age radhlw1 bf7m SigDoseVactnFw1 = no VactnModMw1 = none inthobMedFw1 = age bf4 illw1 bf4m inthobMedMw1 = age inthobMw1 = age InthbModFw1 = none MainEffInthbFw1 = age radhlw1 bf4 SigdoseInthbFw1 = no InthbModMw1 = none MainEffInthbMw1 = age radhlw1 shfamw1 SigDoseInthbMw1 = no MainEffMw1 = radhlw1 bf4 bf40 sxlifeMedMw1 = radhlw1 sxlifeMedFw1 = age illw1 radhlw1 bf4 bf4m MainEffsxlifeFw1 = age bf4 bf4m SigDoseSxlifeFw1 = no SigDosesxlifeMw1 = no MainEffsxlifeMw1 = age bf4 bf40 SigDosePrbfmhmMw1 = no vactnModMw1 = none SigDoseVactnMw1 = no SxLifeModFw1 = no sxlifeModFw1 = none sxlifeModMw1 = none MaineffhmcrMw1 = age bf4 bf40 SigDoseMEhmcrW1 = no PrbfmhmMedFw1 = age bf4 PrbfmhmMedMw1 = age MainEffPrbfmhmFw1 = age radhlw1 bf4 MainEffPrbfmhmMw1 = age bf4 PrbfmhmModFw1 = none PrbfmhmModMw1 = none SigDosePrbfmhmFw1 = no SigDosePrbfhmMw1 = no MainEffPrbfhmMw1 = age bf4 MainEffVactnMw2 = age radhlw2 sxlifeMedMw2 = age illw2 SigDoseSxlifeFw2 = no MainEffsxlifeFw2 = age radhlw2 bf4 bf4m MainEffPrbsocMw2 = age radhlw2 shjobw2 MainEffhmcrFw2 = age hmcrMedFw2 = age bf4 MainEffwkFw2 = age MainEffwkMw2 = age MainEffPrbsocMw1 = age bf4m SigdoseMw1 = no ProbsocMedFw1 = age bf4 ProbsocMedMw1 = radhlw1 ProbsocModFw1 = none SigDoseProbsocMw1 = no hmcrmedMw1 = radhlw1 hmcrmedFw1 = age b4 b40 SigdosehmcrFw1 = no MainEffProbSocFw1 = age radhlw1 avgcumdosew1 shrelaw1 bf4 SigDoseProbsocFw1 = yes PrbsocModMw1 = shjobw1Xd1 shrelaw1Xd1 WkhmcrMw1 = age b4 WkModFw1 = ageXd1 hmcareMedFw1 = age illw1 hmcareMedMw1 = age SigDosehmcrFw1 = no wkMedMw1 = bf40 hmcrModFw1 = none SigDoseHmcrFw1 = yes WkMedMw1 = bf40 hmcrModMw1 = none SigDosehmcrMw1 = no wkMedFw1 = age b4 WKModMw1 = none SigDoseWkMw1 = no SigDoseWkFw1 = no SigDoseFw1 = no wkModFw1 = none wkModMw1 = none VactnMedMw2 = age inthobMedMw2 = age inthobMw2 = age PrbfmhmModMw2 = none MainEffProbSocFw2 = age radhlw2 avgcumdosew2 bf4 hmcrModMw2 = none MainEffhmcrMw2 = age wkMedFw2 = age b4 wkMedMw2 = age bf4 MainEffsxlifeMw2 = age bf4 bf40 shjobw2 shrelaw2 radhlw2 MainEffPrbfmhmMw2 = bf4 bf6 bf7 ProbsocMedFw2 = age bf4 radhlw2 hmcareMedFw2 = age bf4 WkhmcrMw2 = age b4 MainEffhmcrw2 = age hmcrModFw2 = none SigDoseHmcrFw2 = yes NumhmcrModMw2 = none SigDosehmcrMw2 = no SigdosehmcrFw2 = yes hmcrMedMw2 = age ageXillw2 SigDosehmcrFw2 = no MainEffhmcareMw2 = age WkMedMw2 = age ageXillw2 wkMedFw3 = radhlw3 age ageXillw3 bf40 bf4m bf1 VactnMedFw2 = age illw2 radhlw2 VacatnModFw2 = none MainEffVactnFw2 = age radhlw2 bf7m SigDoseVactnFw2 = no VactnModMw2 = none vactnModMw2 = none SigDoseVactnMw2 = no inthobMedFw2 = age bf4 illw2 bf4m InthbModFw2 = none MainEffInthbFw2 = age radhlw2 bf4 SigdoseInthbFw2 = no InthbModMw2 = none MainEffInthbMw2 = age radhlw2 shfamw2 SigDoseInthbMw2 = no MainEffMw2 = radhlw2 bf4 bf40 SigdoseMEinthob = no sxlifeMedFw2 = age illw2 radhlw2 bf4 bf4m SxLifeModFw2 = no sxlifeModFw2 = none sxlifeModMw2 = none SigDosesxlifeMw2 = no PrbfmhmMedFw2 = age bf4 PrbfmhmMedMw2 = age PrbfmhmModFw2 = none MainEffPrbfmhmFw2 = age bf4 bf40 SigDosePrbfmhmFw2 = no PrbfmhmModw2 = none SigDosePrbfmhmMw2 = no SigDosePrbfhmMw2 = no MainEffPrbfhmMw2 = bf4 bf6 bf7 ProbsocMedMw2 = age ProbsocModFw2 = none SigDoseProbsocFw2 = yes ProbSocModMw2 = none SigDoseProbsocMw2 = no PrbsocModMw2 = none SigdoseMw2 = none hmcareMedMw2 = age hmcareModFw2 = none MainEffhmcarew2 = age SigdoseHmcareFw2 = no hmcareModMw2 = none SigDoseHmcareMw2 = no NameMedMw2 = age ageXillw2 NumModMw2 = none SigDosehmcareMw2 = no SigDoseWKMw2 = no WkMedFw2 = age bf4 WkModFw2 = none WKModMw2 = none SigDoseWkMw2 = no SigDoseWkFw2 = no SigDoseFw2 = no wkModFw2 = none wkModMw2 = none VactnMedFw3 = age illw3 radhlw3 VactnMedMw3 = age illw3 VacatnModFw3 = none MainEffVactnFw3 = age radhlw3 deaw3 SigDoseVactnFw3 = no vactnModMw3 = none MainEffVactnMw3 = age bf7m radhlw3 SigDoseVactnMw3 = no sxLifeMedFw3 = age bf4 bf4m sxLifeMedMw3 = age illw3 InthbModFw3 = none MainEffInthbFw3 = age radhlw3 bf4 SigdoseInthbFw3 = no InthbMw3 = none MainEffInthbMw3 = age radhlw3 shfamw3 SigDoseInthbMw3 = no sxlifeMedFw3 = age illw3 radhlw3 bf4 bf4m sxlifeMedMw3 = age illw3 sxlifeModFw3 = none MainEffsxlifeFw3 = age radhlw3 bf4 bf4m shrelaw3 shfamw3 SigDoseSxlifeFw3 = no sxlifeModMw3 = none SigDosesxlifeMw3 = no MainEffsxlifeMw3 = age bf4 illw3 radhlw3 PrbfmhmMedFw3 = age bf4 PrbfmhmMedMw3 = age PrbfmhmModFw3 = none MainEffPrbfmhmFw3 = age bf4 bf40 SigDosePrbfmhmFw3 = no PrbfmhmModw3 = none SigDosePrbfmhmMw3 = no SigDosePrbfhmMw3 = no MainEffPrbfhmMw3 = bf1 bf4 dvcew3 bf7m ProbsocMedFw3 = age radhlw3 ProbsocMedMw3 = age ProbsocModFw3 = none MainEffProbSocFw3 = age radhlw3 illw3 Shrelaw3 avgcumodsew3 SigDoseProbsocFw3 = yes ProbSocModMw3 = none SigDoseProbsocMw3 = no MainEffPrbsocMw3 = age radhlw3 shjobw3 hmcareMedFw3 = age illw3 hmcareMedMw3 = age illw3 hmcareModFw3 = none SigdoseHmcareFw3 = no hmcareModMw3 = none MainEffhmcareMw3 = none SigDoseHmcareMw3 = no wkMedMw3 = bf8 age illw3 ageXillw3 wkModFw3 = none wkModMw3 = none MainEffwkFw3 = age MainEffwkMw3 = workM: age bf8 illw3 shjobw3 SigDoseWKMw3 = no SigDoseWkFw3 = no 752 . 753 . 754 . 755 . *----------- Chunk 5 Dose => Problems with the Family at home Impact 756 . title "4. H1 pt 2 wave 1 Dose = Fam Problems at home impact" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 4. H1 pt 2 wave 1 Dose = Fam Problems at home impact ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:15:56 ***** ******************************************************************************* ******************************************************************************* 757 . forvalues j=1/1 { 2. set more off 3. 758 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 759 . foreach var in HP2pbfhm { 5. forvalues k=1/2 { 6. local w1bf bf1 bf4 bf9 bf10 bf11 bf14 bf15m bf20 bf22 bf30 bf40 7. 760 . di as input "Full main model for `var' for wave= `j' " 8. di _skip(4) 9. di as input "chunk 5 H1 test:Gender= `k' model Wave = `j' for `e(depvar > )' " 10. di _skip(4) 11. 761 . xi: logistic `var' age i.educ occ1w`j'-occ8w`j' /// > marrw`j'1- marrw`j'3 marrw`j'5-marrw`j'6 inc1w`j'-inc4w`j' // > / > radhlw`j' havmil avgcumdosew`j' `w`j'bf' /// > deaw`j' dvcew`j' sepaw`j' accdw`j' movew`j' /// > illw`j' shfamw`j' shhlw`j' shjobw`j' shrelaw`j' suprtw`j' suc > hrw`j' /// > havmilsq if gender==`k', coef difficult iterate(50) 12. estat class 13. estat gof 14. fitstat 15. } 16. } 17. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2pbfhm for wave= 1 chunk 5 H1 test:Gender= 1 model Wave = 1 for HP2probsoc i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: _Ieduc_2 != 0 predicts failure perfectly _Ieduc_2 dropped and 10 obs not used note: _Ieduc_4 != 0 predicts failure perfectly _Ieduc_4 dropped and 12 obs not used note: _Ieduc_7 != 0 predicts failure perfectly _Ieduc_7 dropped and 4 obs not used note: _Ieduc_8 != 0 predicts failure perfectly _Ieduc_8 dropped and 2 obs not used note: occ5w1 != 0 predicts failure perfectly occ5w1 dropped and 8 obs not used note: occ6w1 != 0 predicts failure perfectly occ6w1 dropped and 4 obs not used note: occ7w1 != 0 predicts failure perfectly occ7w1 dropped and 4 obs not used note: marrw12 != 0 predicts failure perfectly marrw12 dropped and 4 obs not used note: marrw15 != 0 predicts success perfectly marrw15 dropped and 1 obs not used note: bf15m != 0 predicts failure perfectly bf15m dropped and 17 obs not used note: dvcew1 != 0 predicts failure perfectly dvcew1 dropped and 1 obs not used note: sepaw1 != 0 predicts failure perfectly sepaw1 dropped and 1 obs not used note: accdw1 != 0 predicts failure perfectly accdw1 dropped and 6 obs not used note: _Ieduc_6 omitted because of collinearity note: marrw16 omitted because of collinearity Logistic regression Number of obs = 259 LR chi2(37) = 58.78 Prob > chi2 = 0.0128 Log likelihood = -41.036775 Pseudo R2 = 0.4173 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1140333 .0507477 2.25 0.025 .0145697 .2134969 _Ieduc_2 | 0 (omitted) _Ieduc_3 | -.0489809 .8274219 -0.06 0.953 -1.670698 1.572736 _Ieduc_4 | 0 (omitted) _Ieduc_5 | -.3717243 1.234713 -0.30 0.763 -2.791717 2.048268 _Ieduc_6 | 0 (omitted) _Ieduc_7 | 0 (omitted) _Ieduc_8 | 0 (omitted) occ1w1 | .0549567 3.40443 0.02 0.987 -6.617603 6.727516 occ2w1 | -.6561172 3.427034 -0.19 0.848 -7.372981 6.060747 occ3w1 | 1.845881 3.621761 0.51 0.610 -5.25264 8.944401 occ4w1 | -.738871 3.594017 -0.21 0.837 -7.783015 6.305273 occ5w1 | 0 (omitted) occ6w1 | 0 (omitted) occ7w1 | 0 (omitted) occ8w1 | .7397449 3.652989 0.20 0.840 -6.419982 7.899471 marrw11 | 9.984281 2273.716 0.00 0.996 -4446.417 4466.386 marrw12 | 0 (omitted) marrw13 | 9.475183 2273.716 0.00 0.997 -4446.927 4465.877 marrw15 | 0 (omitted) marrw16 | 0 (omitted) inc1w1 | -2.098091 3.351876 -0.63 0.531 -8.667648 4.471466 inc2w1 | -1.263123 3.216941 -0.39 0.695 -7.568212 5.041965 inc3w1 | -1.666932 3.248691 -0.51 0.608 -8.034249 4.700384 inc4w1 | -2.798082 3.810066 -0.73 0.463 -10.26567 4.669509 radhlw1 | .0221794 .012356 1.80 0.073 -.0020379 .0463968 havmil | -.0042208 .0093253 -0.45 0.651 -.0224981 .0140566 avgcumdosew1 | -1.041178 .9206409 -1.13 0.258 -2.845601 .7632452 bf1 | -.0255316 .0778527 -0.33 0.743 -.1781201 .127057 bf4 | -.2441414 .1077011 -2.27 0.023 -.4552317 -.0330512 bf9 | -.0547201 .051475 -1.06 0.288 -.1556092 .046169 bf10 | -.2223925 .2030607 -1.10 0.273 -.6203843 .1755992 bf11 | -.0227027 .1366459 -0.17 0.868 -.2905237 .2451183 bf14 | .0001079 .0001172 0.92 0.357 -.0001219 .0003377 bf15m | 0 (omitted) bf20 | .0093433 .0710822 0.13 0.895 -.1299752 .1486619 bf22 | .000374 .0002074 1.80 0.071 -.0000324 .0007805 bf30 | -.0000361 .0007108 -0.05 0.959 -.0014293 .001357 bf40 | -.5322776 .3877192 -1.37 0.170 -1.292193 .227638 deaw1 | .139745 .7195497 0.19 0.846 -1.270547 1.550036 dvcew1 | 0 (omitted) sepaw1 | 0 (omitted) accdw1 | 0 (omitted) movew1 | -.9959846 2.338912 -0.43 0.670 -5.580168 3.588199 illw1 | .0424127 .9714262 0.04 0.965 -1.861548 1.946373 shfamw1 | .0013379 .0148676 0.09 0.928 -.0278022 .0304779 shhlw1 | -.002734 .0255506 -0.11 0.915 -.0528123 .0473444 shjobw1 | -.0385947 .0207074 -1.86 0.062 -.0791804 .0019911 shrelaw1 | .004462 .0134181 0.33 0.739 -.0218371 .0307611 suprtw1 | .0496362 .0609811 0.81 0.416 -.0698846 .169157 suchrw1 | -.0362505 .1014181 -0.36 0.721 -.2350263 .1625254 havmilsq | 4.65e-06 9.99e-06 0.46 0.642 -.0000149 .0000242 _cons | -11.63773 2273.725 -0.01 0.996 -4468.057 4444.781 ------------------------------------------------------------------------------ Note: 8 failures and 0 successes completely determined. Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 9 2 | 11 - | 11 237 | 248 -----------+--------------------------+----------- Total | 20 239 | 259 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 45.00% Specificity Pr( -|~D) 99.16% Positive predictive value Pr( D| +) 81.82% Negative predictive value Pr(~D| -) 95.56% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.84% False - rate for true D Pr( -| D) 55.00% False + rate for classified + Pr(~D| +) 18.18% False - rate for classified - Pr( D| -) 4.44% -------------------------------------------------- Correctly classified 94.98% -------------------------------------------------- Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 259 number of covariate patterns = 259 Pearson chi2(221) = 143.82 Prob > chi2 = 1.0000 Measures of Fit for logistic of HP2pbfhm Log-Lik Intercept Only: -70.429 Log-Lik Full Model: -41.037 D(206): 82.074 LR(37): 58.785 Prob > LR: 0.013 McFadden's R2: 0.417 McFadden's Adj R2: -0.335 Maximum Likelihood R2: 0.203 Cragg & Uhler's R2: 0.484 McKelvey and Zavoina's R2: 0.857 Efron's R2: 0.352 Variance of y*: 23.001 Variance of error: 3.290 Count R2: 0.950 Adj Count R2: 0.350 AIC: 0.726 AIC*n: 188.074 BIC: -1062.633 BIC': 146.818 Full main model for HP2pbfhm for wave= 1 chunk 5 H1 test:Gender= 2 model Wave = 1 for HP2pbfhm i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: marrw12 != 0 predicts failure perfectly marrw12 dropped and 4 obs not used note: bf15m != 0 predicts failure perfectly bf15m dropped and 11 obs not used note: _Ieduc_8 omitted because of collinearity Logistic regression Number of obs = 343 LR chi2(49) = 131.00 Prob > chi2 = 0.0000 Log likelihood = -71.53941 Pseudo R2 = 0.4780 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1418493 .0387097 3.66 0.000 .0659796 .217719 _Ieduc_2 | 17.22542 4534.447 0.00 0.997 -8870.128 8904.579 _Ieduc_3 | 17.37005 4534.447 0.00 0.997 -8869.983 8904.723 _Ieduc_4 | 17.2893 4534.447 0.00 0.997 -8870.064 8904.643 _Ieduc_5 | 17.58061 4534.447 0.00 0.997 -8869.773 8904.934 _Ieduc_6 | 17.54436 4534.447 0.00 0.997 -8869.809 8904.898 _Ieduc_7 | 19.56629 4534.448 0.00 0.997 -8867.789 8906.922 _Ieduc_8 | 0 (omitted) occ1w1 | -2.933908 3.752469 -0.78 0.434 -10.28861 4.420796 occ2w1 | -3.629362 3.805266 -0.95 0.340 -11.08755 3.828822 occ3w1 | -2.149571 3.775043 -0.57 0.569 -9.54852 5.249378 occ4w1 | -3.08403 3.884516 -0.79 0.427 -10.69754 4.529481 occ5w1 | -1.54127 3.940317 -0.39 0.696 -9.264149 6.181609 occ6w1 | -4.518852 4.376561 -1.03 0.302 -13.09675 4.05905 occ7w1 | -.5209646 3.827304 -0.14 0.892 -8.022342 6.980413 occ8w1 | -.6982657 3.788908 -0.18 0.854 -8.124388 6.727857 marrw11 | 6.563183 2.259003 2.91 0.004 2.135619 10.99075 marrw12 | 0 (omitted) marrw13 | 3.664968 1.807756 2.03 0.043 .1218322 7.208105 marrw15 | 4.919489 2.358945 2.09 0.037 .2960419 9.542936 marrw16 | 3.80604 2.237487 1.70 0.089 -.5793546 8.191435 inc1w1 | .8452115 3.777625 0.22 0.823 -6.558798 8.249221 inc2w1 | 2.888908 3.720317 0.78 0.437 -4.402779 10.18059 inc3w1 | 2.900029 3.700306 0.78 0.433 -4.352438 10.1525 inc4w1 | 3.648558 4.001972 0.91 0.362 -4.195162 11.49228 radhlw1 | .0284271 .0105574 2.69 0.007 .0077349 .0491194 havmil | -.0074706 .0174749 -0.43 0.669 -.0417209 .0267796 avgcumdosew1 | -.072348 .4728314 -0.15 0.878 -.9990805 .8543844 bf1 | .0059636 .0469599 0.13 0.899 -.086076 .0980033 bf4 | -.4047147 .0856612 -4.72 0.000 -.5726076 -.2368217 bf9 | -.0595004 .0425833 -1.40 0.162 -.1429621 .0239614 bf10 | -.0062201 .0202319 -0.31 0.759 -.045874 .0334338 bf11 | -.0101918 .0656234 -0.16 0.877 -.1388112 .1184277 bf14 | -.0003249 .0001287 -2.52 0.012 -.0005772 -.0000727 bf15m | 0 (omitted) bf20 | -.0168987 .0398949 -0.42 0.672 -.0950913 .0612938 bf22 | .0003491 .0001303 2.68 0.007 .0000938 .0006044 bf30 | .0000363 .0004332 0.08 0.933 -.0008127 .0008853 bf40 | -.6041257 .200322 -3.02 0.003 -.9967496 -.2115017 deaw1 | -.2106609 .3829692 -0.55 0.582 -.9612668 .539945 dvcew1 | 3.227642 1.990455 1.62 0.105 -.6735776 7.128863 sepaw1 | -1.961135 2.530088 -0.78 0.438 -6.920017 2.997746 accdw1 | -1.15855 1.327999 -0.87 0.383 -3.76138 1.44428 movew1 | .8502008 .7012493 1.21 0.225 -.5242225 2.224624 illw1 | 1.167312 .5375125 2.17 0.030 .1138066 2.220817 shfamw1 | -.0177543 .0110296 -1.61 0.107 -.039372 .0038633 shhlw1 | -.0133242 .0167146 -0.80 0.425 -.0460843 .0194358 shjobw1 | .0113593 .0116917 0.97 0.331 -.0115561 .0342746 shrelaw1 | .0021827 .0102849 0.21 0.832 -.0179754 .0223408 suprtw1 | .0039965 .0137794 0.29 0.772 -.0230105 .0310036 suchrw1 | -.0313303 .0297859 -1.05 0.293 -.0897095 .027049 havmilsq | -5.84e-06 .0000521 -0.11 0.911 -.0001079 .0000962 _cons | -26.74862 4534.45 -0.01 0.995 -8914.107 8860.609 ------------------------------------------------------------------------------ Note: 3 failures and 0 successes completely determined. Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 26 5 | 31 - | 21 291 | 312 -----------+--------------------------+----------- Total | 47 296 | 343 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 55.32% Specificity Pr( -|~D) 98.31% Positive predictive value Pr( D| +) 83.87% Negative predictive value Pr(~D| -) 93.27% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.69% False - rate for true D Pr( -| D) 44.68% False + rate for classified + Pr(~D| +) 16.13% False - rate for classified - Pr( D| -) 6.73% -------------------------------------------------- Correctly classified 92.42% -------------------------------------------------- Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 343 number of covariate patterns = 343 Pearson chi2(293) = 979.77 Prob > chi2 = 0.0000 Measures of Fit for logistic of HP2pbfhm Log-Lik Intercept Only: -137.038 Log-Lik Full Model: -71.539 D(290): 143.079 LR(49): 130.998 Prob > LR: 0.000 McFadden's R2: 0.478 McFadden's Adj R2: 0.091 Maximum Likelihood R2: 0.317 Cragg & Uhler's R2: 0.577 McKelvey and Zavoina's R2: 0.915 Efron's R2: 0.495 Variance of y*: 38.594 Variance of error: 3.290 Count R2: 0.924 Adj Count R2: 0.447 AIC: 0.726 AIC*n: 249.079 BIC: -1549.863 BIC': 155.051 762 . 763 . 764 . title4 "Partly Trimmed male wave 1 Dose => Problems with Family at home mode > ls" ------------------------------------------------------------------------------- Partly Trimmed male wave 1 Dose => Problems with Family at home models ------------------------------------------------------------------------------- 765 . local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 766 . logit HP2pbfhm age bf4 bf22 radhlw1 avgcumdosew1 shjobw1 suprtw1 /// > illw1 if gender==1, iterate(50) Iteration 0: log likelihood = -81.506236 Iteration 1: log likelihood = -70.36767 Iteration 2: log likelihood = -63.735881 Iteration 3: log likelihood = -63.262978 Iteration 4: log likelihood = -63.251084 Iteration 5: log likelihood = -63.251044 Iteration 6: log likelihood = -63.251044 Logistic regression Number of obs = 340 LR chi2(8) = 36.51 Prob > chi2 = 0.0000 Log likelihood = -63.251044 Pseudo R2 = 0.2240 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0417694 .0232436 1.80 0.072 -.0037872 .0873261 bf4 | -.1944454 .0674328 -2.88 0.004 -.3266114 -.0622795 bf22 | .0001979 .0001096 1.81 0.071 -.0000169 .0004126 radhlw1 | .0156873 .0077256 2.03 0.042 .0005455 .0308292 avgcumdosew1 | -.7288118 .6058279 -1.20 0.229 -1.916213 .4585891 shjobw1 | -.0246328 .0097441 -2.53 0.011 -.0437309 -.0055347 suprtw1 | -.0367192 .0154384 -2.38 0.017 -.0669779 -.0064605 illw1 | 1.081626 .5960837 1.81 0.070 -.0866765 2.249929 _cons | -2.68461 1.769319 -1.52 0.129 -6.152411 .7831919 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. 767 . 768 . estat class Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 1 1 | 2 - | 21 317 | 338 -----------+--------------------------+----------- Total | 22 318 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 4.55% Specificity Pr( -|~D) 99.69% Positive predictive value Pr( D| +) 50.00% Negative predictive value Pr(~D| -) 93.79% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.31% False - rate for true D Pr( -| D) 95.45% False + rate for classified + Pr(~D| +) 50.00% False - rate for classified - Pr( D| -) 6.21% -------------------------------------------------- Correctly classified 93.53% -------------------------------------------------- 769 . estat gof Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 340 number of covariate patterns = 337 Pearson chi2(328) = 247.99 Prob > chi2 = 0.9997 770 . fitstat Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -81.506 Log-Lik Full Model: -63.251 D(331): 126.502 LR(8): 36.510 Prob > LR: 0.000 McFadden's R2: 0.224 McFadden's Adj R2: 0.114 Maximum Likelihood R2: 0.102 Cragg & Uhler's R2: 0.267 McKelvey and Zavoina's R2: 0.537 Efron's R2: 0.137 Variance of y*: 7.110 Variance of error: 3.290 Count R2: 0.935 Adj Count R2: 0.000 AIC: 0.425 AIC*n: 144.502 BIC: -1802.879 BIC': 10.121 771 . 772 . title4 "trimmed male main effects wv 1" " Dose => Problems with Family at hom > e models" ------------------------------------------------------------------------------- trimmed male main effects wv 1 ------------------------------------------------------------------------------- 773 . local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 774 . sw, pr(.1):logit HP2pbfhm age sepaw1 dvcew1 radhlw1 avgcumdosew1 bf4 suprtw1 > /// > havmilsq illw1 if gender==1, iterate(50) note: sepaw1 dropped because of estimability note: dvcew1 dropped because of estimability note: o.sepaw1 dropped because of estimability note: o.dvcew1 dropped because of estimability note: 6 obs. dropped because of estimability begin with full model p = 0.7261 >= 0.1000 removing havmilsq p = 0.3270 >= 0.1000 removing avgcumdosew1 p = 0.1493 >= 0.1000 removing radhlw1 p = 0.1420 >= 0.1000 removing suprtw1 p = 0.5106 >= 0.1000 removing illw1 Logistic regression Number of obs = 334 LR chi2(2) = 19.70 Prob > chi2 = 0.0001 Log likelihood = -65.847499 Pseudo R2 = 0.1301 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .043747 .0220586 1.98 0.047 .0005129 .0869811 bf4 | -.1234479 .0425264 -2.90 0.004 -.2067982 -.0400977 _cons | -3.800943 1.442 -2.64 0.008 -6.627212 -.9746754 ------------------------------------------------------------------------------ 775 . 776 . estat class Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 0 0 | 0 - | 20 314 | 334 -----------+--------------------------+----------- Total | 20 314 | 334 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 0.00% Specificity Pr( -|~D) 100.00% Positive predictive value Pr( D| +) .% Negative predictive value Pr(~D| -) 94.01% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.00% False - rate for true D Pr( -| D) 100.00% False + rate for classified + Pr(~D| +) .% False - rate for classified - Pr( D| -) 5.99% -------------------------------------------------- Correctly classified 94.01% -------------------------------------------------- 777 . estat gof Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 334 number of covariate patterns = 219 Pearson chi2(216) = 205.20 Prob > chi2 = 0.6901 778 . fitstat Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -75.697 Log-Lik Full Model: -65.847 D(331): 131.695 LR(2): 19.699 Prob > LR: 0.000 McFadden's R2: 0.130 McFadden's Adj R2: 0.090 Maximum Likelihood R2: 0.057 Cragg & Uhler's R2: 0.157 McKelvey and Zavoina's R2: 0.221 Efron's R2: 0.091 Variance of y*: 4.225 Variance of error: 3.290 Count R2: 0.940 Adj Count R2: 0.000 AIC: 0.412 AIC*n: 137.695 BIC: -1791.793 BIC': -8.077 779 . 780 . scalar MainEffPrbfhmMw1 = "age bf4" 781 . scalar SigDosePrbfhmMw1 = "no" 782 . // construction of moderators for male model 783 . 784 . foreach var in age bf4 { 2. cap gen `var'Xd1 = `var'*avgcumdosew1 3. } 785 . 786 . 787 . 788 . 789 . ***************************************************************************** > * 790 . *-------chunk 6 continued -testing moderators and none found for males 791 . local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 792 . 793 . 794 . title4 "fully Trimmed male main effects wv 1" /// > "Dose => Problems with Family at home models" ------------------------------------------------------------------------------- fully Trimmed male main effects wv 1 ------------------------------------------------------------------------------- 795 . local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 796 . logit HP2pbfhm age radhlw1 avgcumdosew1 /// > bf4Xd1 ageXd1 if /// > gender==1, iterate(50) Iteration 0: log likelihood = -81.506236 Iteration 1: log likelihood = -73.611243 Iteration 2: log likelihood = -71.601053 Iteration 3: log likelihood = -71.266265 Iteration 4: log likelihood = -71.252753 Iteration 5: log likelihood = -71.252739 Iteration 6: log likelihood = -71.252739 Logistic regression Number of obs = 340 LR chi2(5) = 20.51 Prob > chi2 = 0.0010 Log likelihood = -71.252739 Pseudo R2 = 0.1258 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0372929 .0360915 1.03 0.301 -.0334452 .108031 radhlw1 | .0167822 .0070862 2.37 0.018 .0028936 .0306709 avgcumdosew1 | -1.15754 8.877823 -0.13 0.896 -18.55775 16.24267 bf4Xd1 | -.1079495 .1337235 -0.81 0.420 -.3700428 .1541438 ageXd1 | .024333 .13587 0.18 0.858 -.2419673 .2906334 _cons | -5.341428 2.129704 -2.51 0.012 -9.515572 -1.167284 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. 797 . 798 . estat class Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 0 0 | 0 - | 22 318 | 340 -----------+--------------------------+----------- Total | 22 318 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 0.00% Specificity Pr( -|~D) 100.00% Positive predictive value Pr( D| +) .% Negative predictive value Pr(~D| -) 93.53% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.00% False - rate for true D Pr( -| D) 100.00% False + rate for classified + Pr(~D| +) .% False - rate for classified - Pr( D| -) 6.47% -------------------------------------------------- Correctly classified 93.53% -------------------------------------------------- 799 . estat gof Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 340 number of covariate patterns = 328 Pearson chi2(322) = 344.70 Prob > chi2 = 0.1839 800 . fitstat Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -81.506 Log-Lik Full Model: -71.253 D(334): 142.505 LR(5): 20.507 Prob > LR: 0.001 McFadden's R2: 0.126 McFadden's Adj R2: 0.052 Maximum Likelihood R2: 0.059 Cragg & Uhler's R2: 0.154 McKelvey and Zavoina's R2: 0.526 Efron's R2: 0.077 Variance of y*: 6.942 Variance of error: 3.290 Count R2: 0.935 Adj Count R2: 0.000 AIC: 0.454 AIC*n: 154.505 BIC: -1804.362 BIC': 8.638 801 . 802 . scalar SigDosePrbfmhmMw1 = "no" 803 . scalar PrbfmhmModMw1 = "none" 804 . scalar MainEffPrbfmhmMw1= "age bf4" 805 . * 3 main effects signif no main effect for dose for males 806 . 807 . 808 . *-------Chunk 6 continued -testing meditors for females 809 . title4 "Partly Trimmed female wave 1" "Dose => Problems with Family at home > models" ------------------------------------------------------------------------------- Partly Trimmed female wave 1 ------------------------------------------------------------------------------- 810 . local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 811 . logit HP2pbfhm age radhlw1 avgcumdosew1 bf4 if gender==2, iterate(50) Iteration 0: log likelihood = -139.75789 Iteration 1: log likelihood = -110.07066 Iteration 2: log likelihood = -103.23731 Iteration 3: log likelihood = -103.06224 Iteration 4: log likelihood = -103.0617 Iteration 5: log likelihood = -103.0617 Logistic regression Number of obs = 362 LR chi2(4) = 73.39 Prob > chi2 = 0.0000 Log likelihood = -103.0617 Pseudo R2 = 0.2626 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0490304 .0173334 2.83 0.005 .0150575 .0830032 radhlw1 | .0175548 .0055923 3.14 0.002 .0065942 .0285154 avgcumdosew1 | -.0824579 .2858577 -0.29 0.773 -.6427286 .4778129 bf4 | -.1768305 .0371748 -4.76 0.000 -.2496918 -.1039692 _cons | -4.266136 1.122392 -3.80 0.000 -6.465984 -2.066287 ------------------------------------------------------------------------------ 812 . 813 . estat class Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 14 5 | 19 - | 33 310 | 343 -----------+--------------------------+----------- Total | 47 315 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 29.79% Specificity Pr( -|~D) 98.41% Positive predictive value Pr( D| +) 73.68% Negative predictive value Pr(~D| -) 90.38% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.59% False - rate for true D Pr( -| D) 70.21% False + rate for classified + Pr(~D| +) 26.32% False - rate for classified - Pr( D| -) 9.62% -------------------------------------------------- Correctly classified 89.50% -------------------------------------------------- 814 . estat gof Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 362 number of covariate patterns = 360 Pearson chi2(355) = 358.55 Prob > chi2 = 0.4373 815 . fitstat Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -139.758 Log-Lik Full Model: -103.062 D(357): 206.123 LR(4): 73.392 Prob > LR: 0.000 McFadden's R2: 0.263 McFadden's Adj R2: 0.227 Maximum Likelihood R2: 0.184 Cragg & Uhler's R2: 0.341 McKelvey and Zavoina's R2: 0.420 Efron's R2: 0.242 Variance of y*: 5.668 Variance of error: 3.290 Count R2: 0.895 Adj Count R2: 0.191 AIC: 0.597 AIC*n: 216.123 BIC: -1897.194 BIC': -49.826 816 . 817 . scalar SigDosePrbfmhmFw1="no" 818 . 819 . *-------Chunk 6 continued -testing meditors for females 820 . title4 "More partly female Trimmed wave 1" "Dose => Problems with Family at > home models" ------------------------------------------------------------------------------- More partly female Trimmed wave 1 ------------------------------------------------------------------------------- 821 . local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 822 . logit HP2pbfhm age bf4 bf40 if gender==2, iterate(50) Iteration 0: log likelihood = -139.89675 Iteration 1: log likelihood = -112.36921 Iteration 2: log likelihood = -106.24923 Iteration 3: log likelihood = -106.12137 Iteration 4: log likelihood = -106.12091 Iteration 5: log likelihood = -106.12091 Logistic regression Number of obs = 363 LR chi2(3) = 67.55 Prob > chi2 = 0.0000 Log likelihood = -106.12091 Pseudo R2 = 0.2414 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0636233 .017269 3.68 0.000 .0297767 .0974699 bf4 | -.2029758 .0392585 -5.17 0.000 -.2799211 -.1260305 bf40 | -.1942568 .091411 -2.13 0.034 -.373419 -.0150946 _cons | -3.085643 1.113267 -2.77 0.006 -5.267607 -.9036792 ------------------------------------------------------------------------------ 823 . 824 . estat class Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 12 3 | 15 - | 35 313 | 348 -----------+--------------------------+----------- Total | 47 316 | 363 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 25.53% Specificity Pr( -|~D) 99.05% Positive predictive value Pr( D| +) 80.00% Negative predictive value Pr(~D| -) 89.94% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 0.95% False - rate for true D Pr( -| D) 74.47% False + rate for classified + Pr(~D| +) 20.00% False - rate for classified - Pr( D| -) 10.06% -------------------------------------------------- Correctly classified 89.53% -------------------------------------------------- 825 . estat gof Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 363 number of covariate patterns = 341 Pearson chi2(337) = 366.00 Prob > chi2 = 0.1331 826 . fitstat Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -139.897 Log-Lik Full Model: -106.121 D(359): 212.242 LR(3): 67.552 Prob > LR: 0.000 McFadden's R2: 0.241 McFadden's Adj R2: 0.213 Maximum Likelihood R2: 0.170 Cragg & Uhler's R2: 0.316 McKelvey and Zavoina's R2: 0.399 Efron's R2: 0.224 Variance of y*: 5.470 Variance of error: 3.290 Count R2: 0.895 Adj Count R2: 0.191 AIC: 0.607 AIC*n: 220.242 BIC: -1903.849 BIC': -49.868 827 . 828 . scalar SigDosePrbfmhmFw1="no" 829 . scalar MainEffPrbfmhmFw1 = "age radhlw1 bf4" 830 . * 3 significant main effects for females 831 . * no significant main effect for dose 832 . 833 . * constructing moderators 834 . 835 . foreach var in bf4 bf40 { 2. cap gen `var'Xd1 = `var'*avgcumdosew1 3. } 836 . 837 . 838 . title4 "testing female moderator effects: no moderator effects for females" ------------------------------------------------------------------------------- testing female moderator effects: no moderator effects for females ------------------------------------------------------------------------------- 839 . 840 . logit HP2pbfhm age bf4 bf40 ageXd1 bf4Xd1 bf40Xd1 if gender==2, iterate(50) Iteration 0: log likelihood = -139.89675 Iteration 1: log likelihood = -112.27547 Iteration 2: log likelihood = -106.13405 Iteration 3: log likelihood = -105.51351 Iteration 4: log likelihood = -105.50804 Iteration 5: log likelihood = -105.50803 Iteration 6: log likelihood = -105.50803 Logistic regression Number of obs = 363 LR chi2(6) = 68.78 Prob > chi2 = 0.0000 Log likelihood = -105.50803 Pseudo R2 = 0.2458 ------------------------------------------------------------------------------ HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0608384 .0177875 3.42 0.001 .0259756 .0957012 bf4 | -.2060931 .0500059 -4.12 0.000 -.304103 -.1080832 bf40 | -.1074502 .1266208 -0.85 0.396 -.3556225 .140722 ageXd1 | .0103515 .0150265 0.69 0.491 -.0190999 .039803 bf4Xd1 | .0257937 .0948566 0.27 0.786 -.1601219 .2117093 bf40Xd1 | -.2219543 .2551348 -0.87 0.384 -.7220092 .2781007 _cons | -3.197134 1.118642 -2.86 0.004 -5.389632 -1.004635 ------------------------------------------------------------------------------ 841 . estat gof Logistic model for HP2pbfhm, goodness-of-fit test number of observations = 363 number of covariate patterns = 359 Pearson chi2(352) = 359.71 Prob > chi2 = 0.3769 842 . estat class Logistic model for HP2pbfhm -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 13 4 | 17 - | 34 312 | 346 -----------+--------------------------+----------- Total | 47 316 | 363 Classified + if predicted Pr(D) >= .5 True D defined as HP2pbfhm != 0 -------------------------------------------------- Sensitivity Pr( +| D) 27.66% Specificity Pr( -|~D) 98.73% Positive predictive value Pr( D| +) 76.47% Negative predictive value Pr(~D| -) 90.17% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.27% False - rate for true D Pr( -| D) 72.34% False + rate for classified + Pr(~D| +) 23.53% False - rate for classified - Pr( D| -) 9.83% -------------------------------------------------- Correctly classified 89.53% -------------------------------------------------- 843 . fitstat Measures of Fit for logit of HP2pbfhm Log-Lik Intercept Only: -139.897 Log-Lik Full Model: -105.508 D(356): 211.016 LR(6): 68.777 Prob > LR: 0.000 McFadden's R2: 0.246 McFadden's Adj R2: 0.196 Maximum Likelihood R2: 0.173 Cragg & Uhler's R2: 0.321 McKelvey and Zavoina's R2: 0.404 Efron's R2: 0.226 Variance of y*: 5.517 Variance of error: 3.290 Count R2: 0.895 Adj Count R2: 0.191 AIC: 0.620 AIC*n: 225.016 BIC: -1887.391 BIC': -33.411 844 . 845 . scalar PrbfmhmModFw1="none" 846 . 847 . ***************************************************************************** > *** 848 . *---------Chunk 6 continued testing mediating effects for Problems with famil > y 849 . * at home 850 . 851 . 852 . des bf4 bf40 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) 853 . glm age avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 854 . glm HP2pbfhm age if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 163.1004 Iteration 2: deviance = 152.8997 Iteration 3: deviance = 151.9532 Iteration 4: deviance = 151.9347 Iteration 5: deviance = 151.9347 Iteration 6: deviance = 151.9347 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 151.9346518 (1/df) Deviance = .4495108 Pearson = 324.8005072 (1/df) Pearson = .9609482 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1818.249 ------------------------------------------------------------------------------ | EIM HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0296385 .0061334 4.83 0.000 .0176173 .0416597 _cons | -3.073816 .3430721 -8.96 0.000 -3.746225 -2.401407 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 855 . 856 . glm bf4 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1026.9659 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 24.75428 Deviance = 8366.946191 (1/df) Deviance = 24.75428 Pearson = 8366.946191 (1/df) Pearson = 24.75428 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.05274 Log likelihood = -1026.965868 BIC = 6396.763 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.1031788 .161925 -0.64 0.524 -.4205461 .2141884 _cons | 12.54134 .2786337 45.01 0.000 11.99523 13.08746 ------------------------------------------------------------------------------ 857 . glm HP2pbfhm bf4 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 160.0024 Iteration 2: deviance = 148.9095 Iteration 3: deviance = 148.0472 Iteration 4: deviance = 148.0358 Iteration 5: deviance = 148.0358 Iteration 6: deviance = 148.0358 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 148.0358223 (1/df) Deviance = .4379758 Pearson = 318.8927909 (1/df) Pearson = .9434698 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1822.148 ------------------------------------------------------------------------------ | EIM HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.0752171 .0127773 -5.89 0.000 -.1002602 -.0501741 _cons | -.6910419 .1483563 -4.66 0.000 -.9818149 -.400269 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 858 . 859 . * age is a mediating effect for females for Dose=> Problems with family at ho > me 860 . glm age avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 861 . glm HP2pbfhm age if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 252.4902 Iteration 2: deviance = 245.4181 Iteration 3: deviance = 245.1325 Iteration 4: deviance = 245.1321 Iteration 5: deviance = 245.1321 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 245.1320769 (1/df) Deviance = .6790362 Pearson = 382.7456824 (1/df) Pearson = 1.060237 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1882.747 ------------------------------------------------------------------------------ | EIM HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .043836 .0066445 6.60 0.000 .030813 .056859 _cons | -3.477625 .3761287 -9.25 0.000 -4.214824 -2.740426 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 862 . 863 . * bf4 is a mediting effect for females for Dose=> Problems with family at hom > e 864 . glm bf4 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 865 . glm HP2pbfhm bf4 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 240.0481 Iteration 2: deviance = 229.4916 Iteration 3: deviance = 228.6166 Iteration 4: deviance = 228.5986 Iteration 5: deviance = 228.5985 Iteration 6: deviance = 228.5985 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 228.5985175 (1/df) Deviance = .6332369 Pearson = 299.6186696 (1/df) Pearson = .8299686 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1899.281 ------------------------------------------------------------------------------ | EIM HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1229375 .0145262 -8.46 0.000 -.1514083 -.0944667 _cons | -.0739521 .1322299 -0.56 0.576 -.3331179 .1852138 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 866 . 867 . 868 . glm bf6 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1783.6957 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 1091.352 Deviance = 393977.8948 (1/df) Deviance = 1091.352 Pearson = 393977.8948 (1/df) Pearson = 1091.352 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 9.838544 Log likelihood = -1783.695663 BIC = 391850 ------------------------------------------------------------------------------ | OIM bf6 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 4.907676 3.160857 1.55 0.121 -1.287489 11.10284 _cons | 53.46082 2.032533 26.30 0.000 49.47713 57.44452 ------------------------------------------------------------------------------ 869 . glm HP2pbfhm bf6 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 270.845 Iteration 2: deviance = 267.9149 Iteration 3: deviance = 267.8425 Iteration 4: deviance = 267.8424 Iteration 5: deviance = 267.8424 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 267.8424366 (1/df) Deviance = .7419458 Pearson = 362.9777402 (1/df) Pearson = 1.005479 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1860.037 ------------------------------------------------------------------------------ | EIM HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf6 | .0094515 .0024584 3.84 0.000 .0046331 .0142698 _cons | -1.703802 .1741973 -9.78 0.000 -2.045223 -1.362382 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 870 . 871 . 872 . glm bf7 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -902.44388 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 8.497656 Deviance = 3067.65399 (1/df) Deviance = 8.497656 Pearson = 3067.65399 (1/df) Pearson = 8.497656 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 4.983162 Log likelihood = -902.4438834 BIC = 939.7746 ------------------------------------------------------------------------------ | OIM bf7 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.3565771 .2789151 -1.28 0.201 -.9032407 .1900865 _cons | 1.056277 .1793514 5.89 0.000 .7047548 1.407799 ------------------------------------------------------------------------------ 873 . glm HP2pbfhm bf7 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 277.1289 Iteration 2: deviance = 275.4145 Iteration 3: deviance = 275.2733 Iteration 4: deviance = 275.2711 Iteration 5: deviance = 275.2711 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 275.2710641 (1/df) Deviance = .7625237 Pearson = 362.9709778 (1/df) Pearson = 1.00546 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1852.608 ------------------------------------------------------------------------------ | EIM HP2pbfhm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf7 | -.0808371 .0385299 -2.10 0.036 -.1563543 -.0053198 _cons | -1.081139 .0750775 -14.40 0.000 -1.228288 -.9339898 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 874 . 875 . scalar MainEffPrbfhmMw1 = "age bf4" 876 . scalar SigDosePrbfhmMw1 = "no" 877 . scalar SigDosePrbfmhmFw1="no" // fix in others 878 . scalar PrbfmhmModMw1 = "none" 879 . scalar PrbfmhmModFw1="none" 880 . scalar MainEffPrbfmhmMw1= "age bf4" 881 . 882 . scalar MainEffPrbfmhmFw1 = "age radhlw1 bf4" 883 . 884 . scalar PrbfmhmMedMw1 = "age" 885 . scalar PrbfmhmMedFw1 = "age bf4" 886 . * Summary of dose=problems with family at home mediating effects 887 . * males mediators age 1 888 . * females mediators age and BSIsoma rescaled (bf4) 2 889 . 890 . title4 "4. Summary matrix for problems with family at home" ------------------------------------------------------------------------------- 4. Summary matrix for problems with family at home ------------------------------------------------------------------------------- 891 . matrix define prbfamMw1 = J(1,8, 0) 892 . matrix define prbfamFw1 = J(1,8, 0) 893 . matrix colnames prbfamMw1= hypnum ptnum wave gender medsig numMAsig numModsi > g numMed 894 . matrix colnames prbfamFw1= hypnum ptnum wave gender medsig numMAsig > numModsig numMed 895 . matrix define prbfamMw1= (1, 2, 1, 1, 0, 2, 0, 0 ) 896 . matrix define prbfamFw1= (1, 2, 1, 2, 0, 3, 0, 2) 897 . matrix rowname prbfamMw1 = prbfamMw1 898 . matrix rowname prbfamFw1 = prbfamFw1 899 . matlist prbfamMw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- prbfamMw1 | 1 2 1 1 0 2 > 0 0 900 . matlist prbfamFw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- prbfamFw1 | 1 2 1 2 0 3 > 0 2 901 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1 \ hmcrMw1 \ hmcrFw1 \ spMw1 \ > spFw1 \ prbfamMw1 \ prbfamFw1) 902 . 903 . matlist H1pt2w1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 1 > 1 2 r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 2 > 0 2 spMw1 | 1 2 1 1 1 2 > 2 1 spFw1 | 1 2 1 2 0 5 > 0 2 prbfamMw1 | 1 2 1 1 0 2 > 0 0 prbfamFw1 | 1 2 1 2 0 3 > 0 2 904 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 905 . matrix rownames H1pt2w1 = wkMw1 wkFw1 hmcrMw1 hmcrFw1 prbsocMw1 prbsocF > w1 prbfhmMw1 prbfhmFw1 906 . matlist H1pt2w1 | hypnum ptnum wave gender medsig numMAsig > numModsig numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 hmcrMw1 | 1 2 1 1 0 2 > 0 1 hmcrFw1 | 1 2 1 2 0 2 > 0 2 prbsocMw1 | 1 2 1 1 1 2 > 2 1 prbsocFw1 | 1 2 1 2 0 5 > 0 2 prbfhmMw1 | 1 2 1 1 0 2 > 0 0 prbfhmFw1 | 1 2 1 2 0 3 > 0 2 907 . 908 . 909 . 910 . ***************************************************************************** > *** 911 . *-------Chunk 7 Dose==> problems with sex life impact 912 . * Chunk 7 General model for all part 2 of Nottingham Health Profile 913 . 914 . title "5. H1 pt 2 wave 1 part2 H1 dose sex life impact" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 5. H1 pt 2 wave 1 part2 H1 dose sex life impact ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:34 ***** ******************************************************************************* ******************************************************************************* 915 . 916 . forvalues j=1/1 { 2. set more off 3. 917 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 918 . foreach var in HP2sxlife { 5. forvalues k=1/2 { 6. local w1bf bf1 bf4 bf9 bf10 bf11 bf14 bf15m bf20 bf22 bf30 bf40 7. title "Full Nottingham Part 2 subscale models for male & females" /// > "Full main model for `var' for wave= `j' " /// > "chunk 7 H1 test:Gender= `k' model Wave = `j' for `e(depvar)' " 8. di _skip(4) 9. 919 . di _skip(4) 10. 920 . 921 . xi: logistic `var' age i.educ occ1w`j'-occ8w`j' /// > marrw`j'1- marrw`j'3 marrw`j'5-marrw`j'6 inc1w`j'-inc4w`j' // > / > radhlw`j' havmil avgcumdosew`j' `w`j'bf' /// > deaw`j' dvcew`j' sepaw`j' accdw`j' movew`j' /// > illw`j' shfamw`j' shhlw`j' shjobw`j' shrelaw`j' suprtw`j' suc > hrw`j' /// > havmilsq radhlw1 if gender==`k', coef difficult iterate(50) > 11. estat class 12. estat gof 13. fitstat 14. } 15. } 16. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Full Nottingham Part 2 subscale models for male & females ***** ***** Full main model for HP2sxlife for wave= 1 ***** ***** chunk 7 H1 test:Gender= 1 model Wave = 1 for HP2pbfhm ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:34 ***** ******************************************************************************* ******************************************************************************* i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: _Ieduc_4 != 0 predicts failure perfectly _Ieduc_4 dropped and 12 obs not used note: _Ieduc_7 != 0 predicts failure perfectly _Ieduc_7 dropped and 4 obs not used note: marrw12 != 0 predicts failure perfectly marrw12 dropped and 5 obs not used note: marrw15 != 0 predicts success perfectly marrw15 dropped and 1 obs not used note: bf15m != 0 predicts failure perfectly bf15m dropped and 20 obs not used note: dvcew1 != 0 predicts failure perfectly dvcew1 dropped and 2 obs not used note: sepaw1 != 0 predicts success perfectly sepaw1 dropped and 1 obs not used note: _Ieduc_8 omitted because of collinearity note: marrw16 omitted because of collinearity note: radhlw1 omitted because of collinearity Logistic regression Number of obs = 288 LR chi2(43) = 149.61 Prob > chi2 = 0.0000 Log likelihood = -77.748591 Pseudo R2 = 0.4904 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .136472 .0366667 3.72 0.000 .0646066 .2083374 _Ieduc_2 | .9061528 2.373186 0.38 0.703 -3.745206 5.557512 _Ieduc_3 | 1.126364 2.057334 0.55 0.584 -2.905936 5.158664 _Ieduc_4 | 0 (omitted) _Ieduc_5 | .4719974 2.096628 0.23 0.822 -3.637317 4.581312 _Ieduc_6 | 1.043192 2.050556 0.51 0.611 -2.975823 5.062208 _Ieduc_7 | 0 (omitted) _Ieduc_8 | 0 (omitted) occ1w1 | -2.803127 2.35705 -1.19 0.234 -7.422861 1.816606 occ2w1 | -2.898533 2.359872 -1.23 0.219 -7.523797 1.726731 occ3w1 | -3.662112 2.523184 -1.45 0.147 -8.607462 1.283238 occ4w1 | -2.435194 2.370439 -1.03 0.304 -7.08117 2.210782 occ5w1 | -5.879936 2.735652 -2.15 0.032 -11.24172 -.5181554 occ6w1 | -.066721 2.7504 -0.02 0.981 -5.457405 5.323963 occ7w1 | -3.83747 3.050515 -1.26 0.208 -9.81637 2.14143 occ8w1 | -2.95823 2.351973 -1.26 0.208 -7.568013 1.651553 marrw11 | 8.464083 907.837 0.01 0.993 -1770.864 1787.792 marrw12 | 0 (omitted) marrw13 | 6.092151 907.8374 0.01 0.995 -1773.236 1785.421 marrw15 | 0 (omitted) marrw16 | 0 (omitted) inc1w1 | 5.396209 3.075845 1.75 0.079 -.6323357 11.42475 inc2w1 | 5.263232 2.974878 1.77 0.077 -.567422 11.09389 inc3w1 | 4.958244 2.931498 1.69 0.091 -.7873865 10.70387 inc4w1 | 3.505305 3.128247 1.12 0.262 -2.625946 9.636556 radhlw1 | .005437 .0073698 0.74 0.461 -.0090075 .0198816 havmil | -.0008616 .007551 -0.11 0.909 -.0156613 .0139381 avgcumdosew1 | .0450522 .105977 0.43 0.671 -.162659 .2527634 bf1 | .0721133 .0467218 1.54 0.123 -.0194598 .1636864 bf4 | -.3272064 .0723192 -4.52 0.000 -.4689494 -.1854635 bf9 | -.0022701 .0341874 -0.07 0.947 -.0692761 .0647359 bf10 | -.0510813 .0368445 -1.39 0.166 -.1232952 .0211325 bf11 | .2631067 .1226965 2.14 0.032 .0226261 .5035874 bf14 | -.0000418 .0000857 -0.49 0.626 -.0002098 .0001262 bf15m | 0 (omitted) bf20 | -.0698283 .0411461 -1.70 0.090 -.1504731 .0108166 bf22 | .0001486 .0001281 1.16 0.246 -.0001025 .0003997 bf30 | -.0002299 .0004245 -0.54 0.588 -.0010618 .0006021 bf40 | .0950229 .1871851 0.51 0.612 -.2718532 .4618989 deaw1 | .659494 .4055746 1.63 0.104 -.1354176 1.454406 dvcew1 | 0 (omitted) sepaw1 | 0 (omitted) accdw1 | -.8665368 1.711081 -0.51 0.613 -4.220195 2.487121 movew1 | 1.595582 .6025445 2.65 0.008 .4146162 2.776547 illw1 | -.39832 .6674749 -0.60 0.551 -1.706547 .9099067 shfamw1 | .0030017 .0095071 0.32 0.752 -.0156319 .0216353 shhlw1 | .0103775 .0141593 0.73 0.464 -.0173742 .0381292 shjobw1 | -.010105 .0099165 -1.02 0.308 -.029541 .009331 shrelaw1 | .0000546 .0095661 0.01 0.995 -.0186946 .0188037 suprtw1 | .0999707 .0455218 2.20 0.028 .0107496 .1891917 suchrw1 | .0224317 .0297831 0.75 0.451 -.035942 .0808055 havmilsq | -1.68e-06 .0000133 -0.13 0.900 -.0000278 .0000245 radhlw1 | 0 (omitted) _cons | -18.64541 907.8458 -0.02 0.984 -1797.99 1760.7 ------------------------------------------------------------------------------ Note: 2 failures and 0 successes completely determined. Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 43 12 | 55 - | 21 212 | 233 -----------+--------------------------+----------- Total | 64 224 | 288 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 67.19% Specificity Pr( -|~D) 94.64% Positive predictive value Pr( D| +) 78.18% Negative predictive value Pr(~D| -) 90.99% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.36% False - rate for true D Pr( -| D) 32.81% False + rate for classified + Pr(~D| +) 21.82% False - rate for classified - Pr( D| -) 9.01% -------------------------------------------------- Correctly classified 88.54% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 288 number of covariate patterns = 288 Pearson chi2(244) = 242.82 Prob > chi2 = 0.5093 Measures of Fit for logistic of HP2sxlife Log-Lik Intercept Only: -152.555 Log-Lik Full Model: -77.749 D(234): 155.497 LR(43): 149.614 Prob > LR: 0.000 McFadden's R2: 0.490 McFadden's Adj R2: 0.136 Maximum Likelihood R2: 0.405 Cragg & Uhler's R2: 0.620 McKelvey and Zavoina's R2: 0.766 Efron's R2: 0.518 Variance of y*: 14.056 Variance of error: 3.290 Count R2: 0.885 Adj Count R2: 0.484 AIC: 0.915 AIC*n: 263.497 BIC: -1169.636 BIC': 93.894 ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Full Nottingham Part 2 subscale models for male & females ***** ***** Full main model for HP2sxlife for wave= 1 ***** ***** chunk 7 H1 test:Gender= 2 model Wave = 1 for HP2sxlife ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:35 ***** ******************************************************************************* ******************************************************************************* i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: bf15m != 0 predicts failure perfectly bf15m dropped and 11 obs not used note: _Ieduc_8 omitted because of collinearity note: radhlw1 omitted because of collinearity Logistic regression Number of obs = 347 LR chi2(50) = 155.79 Prob > chi2 = 0.0000 Log likelihood = -124.80461 Pseudo R2 = 0.3843 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1290605 .0254171 5.08 0.000 .0792439 .1788771 _Ieduc_2 | -10.67087 962.4723 -0.01 0.991 -1897.082 1875.74 _Ieduc_3 | -9.946262 962.4723 -0.01 0.992 -1896.357 1876.465 _Ieduc_4 | -9.785405 962.4726 -0.01 0.992 -1896.197 1876.626 _Ieduc_5 | -11.49393 962.4724 -0.01 0.990 -1897.905 1874.917 _Ieduc_6 | -10.32295 962.4722 -0.01 0.991 -1896.734 1876.088 _Ieduc_7 | -9.03504 962.4749 -0.01 0.993 -1895.451 1877.381 _Ieduc_8 | 0 (omitted) occ1w1 | -2.600034 2.055585 -1.26 0.206 -6.628907 1.42884 occ2w1 | -1.277677 2.051101 -0.62 0.533 -5.297761 2.742406 occ3w1 | -1.369222 2.057018 -0.67 0.506 -5.400903 2.662459 occ4w1 | -1.559837 2.12733 -0.73 0.463 -5.729327 2.609653 occ5w1 | -2.429826 2.192792 -1.11 0.268 -6.727619 1.867967 occ6w1 | -3.419242 2.325888 -1.47 0.142 -7.977899 1.139414 occ7w1 | -2.504352 2.176283 -1.15 0.250 -6.769788 1.761083 occ8w1 | -1.334187 2.100551 -0.64 0.525 -5.451192 2.782817 marrw11 | -.7193048 1.925081 -0.37 0.709 -4.492394 3.053784 marrw12 | .4856708 2.564006 0.19 0.850 -4.539688 5.51103 marrw13 | -2.16757 1.867036 -1.16 0.246 -5.826893 1.491753 marrw15 | -3.042565 2.275798 -1.34 0.181 -7.503047 1.417918 marrw16 | -1.841276 2.263324 -0.81 0.416 -6.27731 2.594759 inc1w1 | 1.620964 1.976337 0.82 0.412 -2.252585 5.494512 inc2w1 | 1.094487 1.946444 0.56 0.574 -2.720473 4.909448 inc3w1 | .8486289 1.968571 0.43 0.666 -3.009699 4.706957 inc4w1 | 1.499966 2.134492 0.70 0.482 -2.683562 5.683494 radhlw1 | .007034 .0058778 1.20 0.231 -.0044862 .0185542 havmil | -.0000651 .0025878 -0.03 0.980 -.0051371 .0050069 avgcumdosew1 | .599348 .3177091 1.89 0.059 -.0233504 1.222046 bf1 | .0066905 .0314561 0.21 0.832 -.0549623 .0683434 bf4 | -.1521557 .0440097 -3.46 0.001 -.2384131 -.0658983 bf9 | -.0233371 .0265441 -0.88 0.379 -.0753625 .0286883 bf10 | -.0152265 .0136272 -1.12 0.264 -.0419352 .0114823 bf11 | -.0221841 .0445913 -0.50 0.619 -.1095814 .0652131 bf14 | -.0000349 .0000708 -0.49 0.622 -.0001736 .0001039 bf15m | 0 (omitted) bf20 | -.0007465 .0262029 -0.03 0.977 -.0521033 .0506104 bf22 | .0000643 .0000756 0.85 0.395 -.000084 .0002126 bf30 | .0002306 .0003007 0.77 0.443 -.0003588 .00082 bf40 | -.0231935 .1067574 -0.22 0.828 -.232434 .1860471 deaw1 | .0559916 .1861244 0.30 0.764 -.3088055 .4207887 dvcew1 | .4752968 3.360983 0.14 0.888 -6.112108 7.062702 sepaw1 | .3375739 3.57474 0.09 0.925 -6.668788 7.343935 accdw1 | -.4123297 .8114545 -0.51 0.611 -2.002751 1.178092 movew1 | .7332906 .5919892 1.24 0.215 -.4269869 1.893568 illw1 | .1110447 .3242585 0.34 0.732 -.5244903 .7465798 shfamw1 | -.0101274 .0070668 -1.43 0.152 -.0239782 .0037233 shhlw1 | -.0168708 .0106127 -1.59 0.112 -.0376714 .0039298 shjobw1 | .0173829 .0076332 2.28 0.023 .002422 .0323437 shrelaw1 | -.0004942 .0067313 -0.07 0.941 -.0136874 .012699 suprtw1 | .0056448 .0100346 0.56 0.574 -.0140226 .0253123 suchrw1 | -.0012594 .0144683 -0.09 0.931 -.0296168 .0270979 havmilsq | -2.25e-07 1.46e-06 -0.15 0.877 -3.08e-06 2.63e-06 radhlw1 | 0 (omitted) _cons | 5.782713 962.4763 0.01 0.995 -1880.636 1892.202 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 57 20 | 77 - | 37 233 | 270 -----------+--------------------------+----------- Total | 94 253 | 347 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 60.64% Specificity Pr( -|~D) 92.09% Positive predictive value Pr( D| +) 74.03% Negative predictive value Pr(~D| -) 86.30% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 7.91% False - rate for true D Pr( -| D) 39.36% False + rate for classified + Pr(~D| +) 25.97% False - rate for classified - Pr( D| -) 13.70% -------------------------------------------------- Correctly classified 83.57% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 347 number of covariate patterns = 347 Pearson chi2(296) = 311.56 Prob > chi2 = 0.2559 Measures of Fit for logistic of HP2sxlife Log-Lik Intercept Only: -202.698 Log-Lik Full Model: -124.805 D(293): 249.609 LR(50): 155.788 Prob > LR: 0.000 McFadden's R2: 0.384 McFadden's Adj R2: 0.118 Maximum Likelihood R2: 0.362 Cragg & Uhler's R2: 0.525 McKelvey and Zavoina's R2: 0.605 Efron's R2: 0.422 Variance of y*: 8.326 Variance of error: 3.290 Count R2: 0.836 Adj Count R2: 0.394 AIC: 1.031 AIC*n: 357.609 BIC: -1464.243 BIC': 136.679 922 . 923 . *-----Chunk 7 dose3 moderator => sex life impact---------------------------- 924 . title4 "Chunk 7 partly trimmed male model of dose=>HP2sxlife relationship in > wave 1" ------------------------------------------------------------------------------- Chunk 7 partly trimmed male model of dose=>HP2sxlife relationship in wave 1 ------------------------------------------------------------------------------- 925 . forvalues j=1/1 { 2. set more off 3. di as input "trimmed HP2sexlife main effects models wave 1 for H1 part 2 > with dose ns" 4. di as input "wave 1 male dose avgcumdosew`j' main effect not signif" 5. logit HP2sxlife age occ1w1-occ7w1 inc1w1-inc4w1 bf4 bf40 /// > if gender==1 6. estat class 7. estat gof 8. fitstat 9. } trimmed HP2sexlife main effects models wave 1 for H1 part 2 with dose ns wave 1 male dose avgcumdosew1 main effect not signif Iteration 0: log likelihood = -171.51396 Iteration 1: log likelihood = -118.07472 Iteration 2: log likelihood = -110.61729 Iteration 3: log likelihood = -110.3575 Iteration 4: log likelihood = -110.35718 Iteration 5: log likelihood = -110.35718 Logistic regression Number of obs = 340 LR chi2(14) = 122.31 Prob > chi2 = 0.0000 Log likelihood = -110.35718 Pseudo R2 = 0.3566 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0941116 .0221002 4.26 0.000 .0507959 .1374272 occ1w1 | -.8108243 .6947007 -1.17 0.243 -2.172413 .5507641 occ2w1 | -.8508826 .7171286 -1.19 0.235 -2.256429 .5546637 occ3w1 | -.8480358 .9624232 -0.88 0.378 -2.734351 1.038279 occ4w1 | -.4031164 .7298235 -0.55 0.581 -1.833544 1.027311 occ5w1 | -3.36185 1.50629 -2.23 0.026 -6.314124 -.4095757 occ6w1 | -.2905401 1.302685 -0.22 0.824 -2.843755 2.262675 occ7w1 | -.2498895 1.738151 -0.14 0.886 -3.656603 3.156824 inc1w1 | 1.792574 .9507668 1.89 0.059 -.0708952 3.656042 inc2w1 | 1.897351 .9299892 2.04 0.041 .0746052 3.720096 inc3w1 | 1.708189 .9374578 1.82 0.068 -.1291946 3.545573 inc4w1 | 1.397502 1.070212 1.31 0.192 -.7000754 3.49508 bf4 | -.2058489 .0355202 -5.80 0.000 -.2754671 -.1362307 bf40 | .2384671 .1033522 2.31 0.021 .0359004 .4410337 _cons | -5.837351 1.560717 -3.74 0.000 -8.896301 -2.778402 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 35 17 | 52 - | 34 254 | 288 -----------+--------------------------+----------- Total | 69 271 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 50.72% Specificity Pr( -|~D) 93.73% Positive predictive value Pr( D| +) 67.31% Negative predictive value Pr(~D| -) 88.19% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 6.27% False - rate for true D Pr( -| D) 49.28% False + rate for classified + Pr(~D| +) 32.69% False - rate for classified - Pr( D| -) 11.81% -------------------------------------------------- Correctly classified 85.00% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 340 number of covariate patterns = 324 Pearson chi2(309) = 279.35 Prob > chi2 = 0.8861 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -171.514 Log-Lik Full Model: -110.357 D(325): 220.714 LR(14): 122.314 Prob > LR: 0.000 McFadden's R2: 0.357 McFadden's Adj R2: 0.269 Maximum Likelihood R2: 0.302 Cragg & Uhler's R2: 0.476 McKelvey and Zavoina's R2: 0.514 Efron's R2: 0.374 Variance of y*: 6.767 Variance of error: 3.290 Count R2: 0.850 Adj Count R2: 0.261 AIC: 0.737 AIC*n: 250.714 BIC: -1673.693 BIC': -40.708 926 . 927 . scalar MaineffhmcrMw1 = "age bf4 bf40" 928 . 929 . title "Chunk 7 trimmed male model of dose and HP2sxlife relationship in wave > 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** *****Chunk 7 trimmed male model of dose and HP2sxlife relationship in wave 1*** > ** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:39 ***** ******************************************************************************* ******************************************************************************* 930 . title4 "h1 pt 2 wave 1 dose=> sex life impact on males" ------------------------------------------------------------------------------- h1 pt 2 wave 1 dose=> sex life impact on males ------------------------------------------------------------------------------- 931 . forvalues j=1/1 { 2. set more off 3. des occ5w1 inc1w1 inc2w1 inc3w1 4. di as input "fully trimmed HP2sexlife main effects models wave 1 for H1 p > art 2 with dose ns" 5. di as input "wave 1 male dose avgcumdosew`j' main effect not signif" 6. logit HP2sxlife age bf4 bf40 shjobw`j' shrelaw`j' radhlw`j' if > gender==1 7. estat class 8. estat gof 9. fitstat 10. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- occ5w1 double %15.0g LABJ factory laborer machinist transp cleaner in 1986 inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 fully trimmed HP2sexlife main effects models wave 1 for H1 part 2 with dose ns wave 1 male dose avgcumdosew1 main effect not signif Iteration 0: log likelihood = -171.51396 Iteration 1: log likelihood = -121.19042 Iteration 2: log likelihood = -114.60006 Iteration 3: log likelihood = -114.41503 Iteration 4: log likelihood = -114.41445 Iteration 5: log likelihood = -114.41445 Logistic regression Number of obs = 340 LR chi2(6) = 114.20 Prob > chi2 = 0.0000 Log likelihood = -114.41445 Pseudo R2 = 0.3329 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0617919 .0157391 3.93 0.000 .0309439 .09264 bf4 | -.1966537 .0409434 -4.80 0.000 -.2769013 -.1164061 bf40 | .2838331 .1042379 2.72 0.006 .0795305 .4881356 shjobw1 | -.0050343 .0063951 -0.79 0.431 -.0175685 .0074999 shrelaw1 | -.0020982 .0048927 -0.43 0.668 -.0116877 .0074913 radhlw1 | .0066138 .0047502 1.39 0.164 -.0026965 .0159241 _cons | -3.247623 1.168417 -2.78 0.005 -5.537679 -.9575672 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 33 16 | 49 - | 36 255 | 291 -----------+--------------------------+----------- Total | 69 271 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 47.83% Specificity Pr( -|~D) 94.10% Positive predictive value Pr( D| +) 67.35% Negative predictive value Pr(~D| -) 87.63% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.90% False - rate for true D Pr( -| D) 52.17% False + rate for classified + Pr(~D| +) 32.65% False - rate for classified - Pr( D| -) 12.37% -------------------------------------------------- Correctly classified 84.71% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 340 number of covariate patterns = 335 Pearson chi2(328) = 293.15 Prob > chi2 = 0.9171 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -171.514 Log-Lik Full Model: -114.414 D(333): 228.829 LR(6): 114.199 Prob > LR: 0.000 McFadden's R2: 0.333 McFadden's Adj R2: 0.292 Maximum Likelihood R2: 0.285 Cragg & Uhler's R2: 0.449 McKelvey and Zavoina's R2: 0.477 Efron's R2: 0.350 Variance of y*: 6.291 Variance of error: 3.290 Count R2: 0.847 Adj Count R2: 0.246 AIC: 0.714 AIC*n: 242.829 BIC: -1712.210 BIC': -79.225 932 . 933 . des bf4 bf40 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) 934 . scalar MainEffsxlifeMw1 = "age bf4 bf40 " 935 . scalar SigDosesxlifeMw1 = "no" 936 . 937 . 938 . forvalues j=1/1 { 2. title4 "trimmed HP2sxlife main effects models wave `j' for H1 part 2 with > dose ns" 3. title4 "Wave `j dose HPsxlife relationship but avgcumdosew`j': Dose not si > gnif" 4. } ------------------------------------------------------------------------------- trimmed HP2sxlife main effects models wave 1 for H1 part 2 with dose ns ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- Wave `j dose HPsxlife relationship but avgcumdosew1: Dose not signif ------------------------------------------------------------------------------- 939 . 940 . cap gen bf4Xd1= bf4*avgcumdosew1 941 . cap gen radhlw1Xd1 = radhlw1*avgcumdosew1 942 . cap gen ageXd1 = age*avgcumdosew1 943 . 944 . set more off 945 . des bf4 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) 946 . forvalues j=1/1 { 2. sw, pr(.1):logistic HP2sxlife age bf4 /// > avgcumdosew`j' ageXd1 bf4Xd1 radhlw1Xd1 if gender==1, coef 3. estat class 4. estat gof 5. fitstat 6. } begin with full model p = 0.6350 >= 0.1000 removing bf4Xd1 p = 0.4431 >= 0.1000 removing radhlw1Xd1 p = 0.1068 >= 0.1000 removing avgcumdosew1 p = 0.4455 >= 0.1000 removing ageXd1 Logistic regression Number of obs = 340 LR chi2(2) = 105.30 Prob > chi2 = 0.0000 Log likelihood = -118.86164 Pseudo R2 = 0.3070 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0735516 .0151718 4.85 0.000 .0438154 .1032879 bf4 | -.2011944 .0322322 -6.24 0.000 -.2643683 -.1380205 _cons | -3.058766 .9461336 -3.23 0.001 -4.913154 -1.204378 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 32 14 | 46 - | 37 257 | 294 -----------+--------------------------+----------- Total | 69 271 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 46.38% Specificity Pr( -|~D) 94.83% Positive predictive value Pr( D| +) 69.57% Negative predictive value Pr(~D| -) 87.41% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 5.17% False - rate for true D Pr( -| D) 53.62% False + rate for classified + Pr(~D| +) 30.43% False - rate for classified - Pr( D| -) 12.59% -------------------------------------------------- Correctly classified 85.00% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 340 number of covariate patterns = 222 Pearson chi2(219) = 215.06 Prob > chi2 = 0.5625 Measures of Fit for logistic of HP2sxlife Log-Lik Intercept Only: -171.514 Log-Lik Full Model: -118.862 D(337): 237.723 LR(2): 105.305 Prob > LR: 0.000 McFadden's R2: 0.307 McFadden's Adj R2: 0.289 Maximum Likelihood R2: 0.266 Cragg & Uhler's R2: 0.419 McKelvey and Zavoina's R2: 0.435 Efron's R2: 0.320 Variance of y*: 5.825 Variance of error: 3.290 Count R2: 0.850 Adj Count R2: 0.261 AIC: 0.717 AIC*n: 243.723 BIC: -1726.631 BIC': -93.647 947 . 948 . scalar sxlifeModMw1 = "none" 949 . *xx male moderators: no main significant dose effect 950 . *xx no male moderators for sexlife impact 951 . 952 . 953 . 954 . title4 "H1 pt2 wave 1 female dose=> sexlife impact models" ------------------------------------------------------------------------------- H1 pt2 wave 1 female dose=> sexlife impact models ------------------------------------------------------------------------------- 955 . 956 . *-----Chunk 7 dose3 moderator => sex life impact---------------------------- 957 . di as input "chunk 7 female wave=3" chunk 7 female wave=3 958 . title "Chunk 7 trimmed female model:" "dose and HP2sxlife relationship in wav > e 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Chunk 7 trimmed female model: ***** ***** dose and HP2sxlife relationship in wave 1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:49 ***** ******************************************************************************* ******************************************************************************* 959 . * female models 960 . forvalues j=1/1 { 2. 961 . set more off 3. des bf4 bf4m shfamw1 shrelaw1 avgcumdosew1 4. title4 "trimmed HP2sexlife main effects models" "wave 1 for H1 part 2 with > dose ns" 5. title4 "wave 1 dose HP2sexlife relationship" "avgcumdosew`j' Dose not sign > if" 6. logit HP2sxlife age radhlw`j' bf4 bf4m /// > avgcumdosew`j' if gender==2 7. estat class 8. estat gof 9. fitstat 10. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 ------------------------------------------------------------------------------- trimmed HP2sexlife main effects models ------------------------------------------------------------------------------- ------------------------------------------------------------------------------- wave 1 dose HP2sexlife relationship ------------------------------------------------------------------------------- Iteration 0: log likelihood = -207.32098 Iteration 1: log likelihood = -149.84784 Iteration 2: log likelihood = -145.01939 Iteration 3: log likelihood = -144.95408 Iteration 4: log likelihood = -144.95402 Iteration 5: log likelihood = -144.95402 Logistic regression Number of obs = 362 LR chi2(5) = 124.73 Prob > chi2 = 0.0000 Log likelihood = -144.95402 Pseudo R2 = 0.3008 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0771236 .0150604 5.12 0.000 .0476059 .1066414 radhlw1 | .0062168 .0040945 1.52 0.129 -.0018083 .0142418 bf4 | -.5739615 .185844 -3.09 0.002 -.9382091 -.2097139 bf4m | .3919975 .1717276 2.28 0.022 .0554176 .7285774 avgcumdosew1 | .3818052 .2467519 1.55 0.122 -.1018197 .8654301 _cons | -7.066627 1.552843 -4.55 0.000 -10.11014 -4.023111 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 51 22 | 73 - | 43 246 | 289 -----------+--------------------------+----------- Total | 94 268 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 54.26% Specificity Pr( -|~D) 91.79% Positive predictive value Pr( D| +) 69.86% Negative predictive value Pr(~D| -) 85.12% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 8.21% False - rate for true D Pr( -| D) 45.74% False + rate for classified + Pr(~D| +) 30.14% False - rate for classified - Pr( D| -) 14.88% -------------------------------------------------- Correctly classified 82.04% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 362 number of covariate patterns = 360 Pearson chi2(354) = 377.83 Prob > chi2 = 0.1837 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -207.321 Log-Lik Full Model: -144.954 D(356): 289.908 LR(5): 124.734 Prob > LR: 0.000 McFadden's R2: 0.301 McFadden's Adj R2: 0.272 Maximum Likelihood R2: 0.291 Cragg & Uhler's R2: 0.427 McKelvey and Zavoina's R2: 0.451 Efron's R2: 0.348 Variance of y*: 5.995 Variance of error: 3.290 Count R2: 0.820 Adj Count R2: 0.309 AIC: 0.834 AIC*n: 301.908 BIC: -1807.517 BIC': -95.276 962 . scalar SigDoseSxlifeFw1 = "no" 963 . scalar MainEffsxlifeFw1 = "age bf4 bf4m" 964 . *----- constructing possible moderators 965 . 966 . foreach var in bf4 bf4m shfamw1 shrelaw1 radhlw1 { 2. cap gen `var'Xd1 = `var'*avgcumdosew1 3. } 967 . 968 . scalar sxlifeModFw1="none" 969 . scalar SigDoseSxlifeFw1 = "none" 970 . 971 . 972 . 973 . *----- testing female moderators 974 . title "partly trimmed female moderator model of dose & HP2sxlife relationship > in wv 1" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** *****partly trimmed female moderator model of dose & HP2sxlife relationship in > wv 1***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:50 ***** ******************************************************************************* ******************************************************************************* 975 . * male models 976 . forvalues j=1/1 { 2. set more off 3. des bf4 bf4m shfamw1 shrelaw1 avgcumdosew1 4. title3 "trimmed HP2sexlife main effects models wave 1 for H1 part 2 with d > ose ns" 5. title "wave 1 dose HP2sexlife relationship but avgcumdosew`j': Dose not si > gnif" 6. logit HP2sxlife age radhlw`j' bf4 bf4m /// > shrelaw`j' avgcumdosew`j' radhlw`j'Xd1 /// > bf4Xd1 shrelaw1Xd1 if gender==2 7. estat class 8. estat gof 9. fitstat 10. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 ------------------------------------------------------------------------------- title3 : trimmed HP2sexlife main effects models wave 1 for H1 part 2 with dose > ns 18 Jun 2012 18:16:50 computer Macintosh (Intel 64-bit) folder /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/ > h1pt2 Data file chwide16june2012.dta currrently has 2386 variables and 703 obse > rvations ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** *****wave 1 dose HP2sexlife relationship but avgcumdosew1: Dose not signif***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:50 ***** ******************************************************************************* ******************************************************************************* Iteration 0: log likelihood = -207.32098 Iteration 1: log likelihood = -147.41511 Iteration 2: log likelihood = -141.75413 Iteration 3: log likelihood = -141.62542 Iteration 4: log likelihood = -141.62494 Iteration 5: log likelihood = -141.62494 Logistic regression Number of obs = 362 LR chi2(9) = 131.39 Prob > chi2 = 0.0000 Log likelihood = -141.62494 Pseudo R2 = 0.3169 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0814535 .0153556 5.30 0.000 .0513571 .11155 radhlw1 | .0024977 .0049558 0.50 0.614 -.0072155 .0122108 bf4 | -.5313372 .1875606 -2.83 0.005 -.8989493 -.1637251 bf4m | .4140426 .1747902 2.37 0.018 .0714601 .7566252 shrelaw1 | -.0104094 .0067432 -1.54 0.123 -.0236259 .0028071 avgcumdosew1 | .8771287 .9794814 0.90 0.371 -1.04262 2.796877 radhlw1Xd1 | .0106434 .0087424 1.22 0.223 -.0064914 .0277781 bf4Xd1 | -.2197957 .1330334 -1.65 0.098 -.4805363 .040945 shrelaw1Xd1 | .0233442 .0154714 1.51 0.131 -.0069791 .0536676 _cons | -7.515707 1.659151 -4.53 0.000 -10.76758 -4.26383 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 51 21 | 72 - | 43 247 | 290 -----------+--------------------------+----------- Total | 94 268 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 54.26% Specificity Pr( -|~D) 92.16% Positive predictive value Pr( D| +) 70.83% Negative predictive value Pr(~D| -) 85.17% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 7.84% False - rate for true D Pr( -| D) 45.74% False + rate for classified + Pr(~D| +) 29.17% False - rate for classified - Pr( D| -) 14.83% -------------------------------------------------- Correctly classified 82.32% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 362 number of covariate patterns = 361 Pearson chi2(351) = 368.67 Prob > chi2 = 0.2479 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -207.321 Log-Lik Full Model: -141.625 D(352): 283.250 LR(9): 131.392 Prob > LR: 0.000 McFadden's R2: 0.317 McFadden's Adj R2: 0.269 Maximum Likelihood R2: 0.304 Cragg & Uhler's R2: 0.446 McKelvey and Zavoina's R2: 0.536 Efron's R2: 0.365 Variance of y*: 7.085 Variance of error: 3.290 Count R2: 0.823 Adj Count R2: 0.319 AIC: 0.838 AIC*n: 303.250 BIC: -1790.609 BIC': -78.367 977 . // trimming further 978 . title "fully female moderator model of dose & HP2sxlife relationship in wv 1" > ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** *****fully female moderator model of dose & HP2sxlife relationship in wv 1***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:51 ***** ******************************************************************************* ******************************************************************************* 979 . 980 . 981 . * female models 982 . set more off 983 . forvalues j=1/1 { 2. des bf4 bf4m shfamw1 shrelaw1 avgcumdosew1 3. title3 "trimmed HP2sexlife main effects models wave 1 for H1 part 2 with d > ose ns" 4. title "wave 1 dose HP2sexlife relationship but avgcumdosew`j': Dose not si > gnif" 5. logit HP2sxlife age radhlw`j' bf4 bf4m /// > shrelaw1 avgcumdosew`j' /// > shrelaw1Xd1 if gender==2 6. estat class 7. estat gof 8. fitstat 9. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 ------------------------------------------------------------------------------- title3 : trimmed HP2sexlife main effects models wave 1 for H1 part 2 with dose > ns 18 Jun 2012 18:16:51 computer Macintosh (Intel 64-bit) folder /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/ > h1pt2 Data file chwide16june2012.dta currrently has 2386 variables and 703 obse > rvations ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** *****wave 1 dose HP2sexlife relationship but avgcumdosew1: Dose not signif***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:51 ***** ******************************************************************************* ******************************************************************************* Iteration 0: log likelihood = -207.32098 Iteration 1: log likelihood = -148.73094 Iteration 2: log likelihood = -143.96367 Iteration 3: log likelihood = -143.91142 Iteration 4: log likelihood = -143.91139 Iteration 5: log likelihood = -143.91139 Logistic regression Number of obs = 362 LR chi2(7) = 126.82 Prob > chi2 = 0.0000 Log likelihood = -143.91139 Pseudo R2 = 0.3059 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0780855 .0151068 5.17 0.000 .0484767 .1076944 radhlw1 | .0057692 .0041216 1.40 0.162 -.0023091 .0138474 bf4 | -.5774463 .1888057 -3.06 0.002 -.9474987 -.2073939 bf4m | .3915506 .1745706 2.24 0.025 .0493985 .7337026 shrelaw1 | -.0078788 .0061361 -1.28 0.199 -.0199053 .0041477 avgcumdosew1 | .0233678 .3583816 0.07 0.948 -.6790473 .7257829 shrelaw1Xd1 | .0143198 .0116222 1.23 0.218 -.0084592 .0370989 _cons | -6.851526 1.575045 -4.35 0.000 -9.938557 -3.764495 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 53 23 | 76 - | 41 245 | 286 -----------+--------------------------+----------- Total | 94 268 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 56.38% Specificity Pr( -|~D) 91.42% Positive predictive value Pr( D| +) 69.74% Negative predictive value Pr(~D| -) 85.66% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 8.58% False - rate for true D Pr( -| D) 43.62% False + rate for classified + Pr(~D| +) 30.26% False - rate for classified - Pr( D| -) 14.34% -------------------------------------------------- Correctly classified 82.32% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 362 number of covariate patterns = 361 Pearson chi2(353) = 375.36 Prob > chi2 = 0.1978 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -207.321 Log-Lik Full Model: -143.911 D(354): 287.823 LR(7): 126.819 Prob > LR: 0.000 McFadden's R2: 0.306 McFadden's Adj R2: 0.267 Maximum Likelihood R2: 0.296 Cragg & Uhler's R2: 0.433 McKelvey and Zavoina's R2: 0.465 Efron's R2: 0.353 Variance of y*: 6.146 Variance of error: 3.290 Count R2: 0.823 Adj Count R2: 0.319 AIC: 0.839 AIC*n: 303.823 BIC: -1797.819 BIC': -85.578 984 . * female models 985 . forvalues j=1/1 { 2. des bf4 bf4m shfamw1 shrelaw1 avgcumdosew1 3. title3 "trimmed HP2sexlife main effects models wave 1 for H1 part 2 with d > ose ns" 4. title "wave 1 dose HP2sexlife relationship but avgcumdosew`j': Dose not si > gnif" 5. logit HP2sxlife age radhlw`j' bf4 bf4m /// > shrelaw1 shfamw`j' /// > if gender==2 6. estat class 7. estat gof 8. fitstat 9. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 avgcumdosew1 double %8.0g wave 1 avg mean CS137 dose in mGy ending 12/31/1986 ------------------------------------------------------------------------------- title3 : trimmed HP2sexlife main effects models wave 1 for H1 part 2 with dose > ns 18 Jun 2012 18:16:53 computer Macintosh (Intel 64-bit) folder /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/ > h1pt2 Data file chwide16june2012.dta currrently has 2386 variables and 703 obse > rvations ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** *****wave 1 dose HP2sexlife relationship but avgcumdosew1: Dose not signif***** ***** ***** ***** ***** ***** 18 Jun 2012 18:16:53 ***** ******************************************************************************* ******************************************************************************* Iteration 0: log likelihood = -207.32098 Iteration 1: log likelihood = -151.03193 Iteration 2: log likelihood = -146.24555 Iteration 3: log likelihood = -146.19381 Iteration 4: log likelihood = -146.19378 Iteration 5: log likelihood = -146.19378 Logistic regression Number of obs = 362 LR chi2(6) = 122.25 Prob > chi2 = 0.0000 Log likelihood = -146.19378 Pseudo R2 = 0.2948 ------------------------------------------------------------------------------ HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0785876 .0150203 5.23 0.000 .0491483 .1080268 radhlw1 | .0058723 .004062 1.45 0.148 -.0020891 .0138337 bf4 | -.5874305 .1868633 -3.14 0.002 -.9536758 -.2211852 bf4m | .3963155 .172212 2.30 0.021 .0587862 .7338448 shrelaw1 | -.0014364 .0051848 -0.28 0.782 -.0115983 .0087256 shfamw1 | -.0015628 .0050062 -0.31 0.755 -.0113747 .0082491 _cons | -6.829393 1.564771 -4.36 0.000 -9.896289 -3.762497 ------------------------------------------------------------------------------ Logistic model for HP2sxlife -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 50 22 | 72 - | 44 246 | 290 -----------+--------------------------+----------- Total | 94 268 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2sxlife != 0 -------------------------------------------------- Sensitivity Pr( +| D) 53.19% Specificity Pr( -|~D) 91.79% Positive predictive value Pr( D| +) 69.44% Negative predictive value Pr(~D| -) 84.83% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 8.21% False - rate for true D Pr( -| D) 46.81% False + rate for classified + Pr(~D| +) 30.56% False - rate for classified - Pr( D| -) 15.17% -------------------------------------------------- Correctly classified 81.77% -------------------------------------------------- Logistic model for HP2sxlife, goodness-of-fit test number of observations = 362 number of covariate patterns = 360 Pearson chi2(353) = 374.57 Prob > chi2 = 0.2058 Measures of Fit for logit of HP2sxlife Log-Lik Intercept Only: -207.321 Log-Lik Full Model: -146.194 D(355): 292.388 LR(6): 122.254 Prob > LR: 0.000 McFadden's R2: 0.295 McFadden's Adj R2: 0.261 Maximum Likelihood R2: 0.287 Cragg & Uhler's R2: 0.420 McKelvey and Zavoina's R2: 0.446 Efron's R2: 0.339 Variance of y*: 5.936 Variance of error: 3.290 Count R2: 0.818 Adj Count R2: 0.298 AIC: 0.846 AIC*n: 306.388 BIC: -1799.146 BIC': -86.905 986 . 987 . scalar MainEffsxlifeFw1 = "age radhlw1 bf4 bf4m" 988 . scalar SigDoseSxlifeFw1="no" 989 . scalar SxLifeModFw1 = "no" 990 . * xx female main effects model: no sign dose main effect 991 . * xx 6 signif main effects 992 . * xx no moderator effects significant 993 . 994 . title4 "h1 pt 2 wave 1 dose-> sexlife sexlife mediator impact models" ------------------------------------------------------------------------------- h1 pt 2 wave 1 dose-> sexlife sexlife mediator impact models ------------------------------------------------------------------------------- 995 . 996 . di as input "testing possible sex life mediator effects for males" testing possible sex life mediator effects for males 997 . 998 . * age is a mediating effect for males for Dose=> sex life for men 999 . des bf4 bf4m storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) 1000 . glm age avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 1001 . glm HP2sxlife age if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 287.05 Iteration 2: deviance = 282.9134 Iteration 3: deviance = 282.8157 Iteration 4: deviance = 282.8156 Iteration 5: deviance = 282.8156 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 282.8156218 (1/df) Deviance = .8367326 Pearson = 327.5643783 (1/df) Pearson = .9691254 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1687.368 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0532907 .0067666 7.88 0.000 .0400283 .0665531 _cons | -3.620299 .3741324 -9.68 0.000 -4.353585 -2.887013 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1002 . 1003 . 1004 . des illw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- illw1 double %8.0g Total number of illnesses experienced in time period 1976-1986 1005 . glm illw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -151.48261 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = .1435742 Deviance = 48.52808533 (1/df) Deviance = .1435742 Pearson = 48.52808533 (1/df) Pearson = .1435742 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = .9028389 Log likelihood = -151.4826069 BIC = -1921.656 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0087277 .0123318 0.71 0.479 -.0154423 .0328976 _cons | .0962541 .0212201 4.54 0.000 .0546635 .1378446 ------------------------------------------------------------------------------ 1006 . glm HP2sxlife illw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 342.7406 Iteration 2: deviance = 342.5208 Iteration 3: deviance = 342.5207 Iteration 4: deviance = 342.5207 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 342.5207329 (1/df) Deviance = 1.013375 Pearson = 340.1346783 (1/df) Pearson = 1.006316 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1627.663 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .1374179 .1941211 0.71 0.479 -.2430524 .5178882 _cons | -.8459321 .0807352 -10.48 0.000 -1.00417 -.6876941 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1007 . 1008 . des radhlw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- radhlw1 double %8.0g how much believed personal health is affected by radiation in 1986 1009 . glm radhlw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1710.3417 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1378.645 Deviance = 465981.8893 (1/df) Deviance = 1378.645 Pearson = 465981.8893 (1/df) Pearson = 1378.645 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.0726 Log likelihood = -1710.341694 BIC = 464011.7 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 2.398285 1.208412 1.98 0.047 .0298407 4.766729 _cons | 44.66477 2.079384 21.48 0.000 40.58925 48.74029 ------------------------------------------------------------------------------ 1010 . glm HP2sxlife radhlw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 314.3547 Iteration 2: deviance = 313.2652 Iteration 3: deviance = 313.2621 Iteration 4: deviance = 313.2621 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 313.2620862 (1/df) Deviance = .9268109 Pearson = 338.905649 (1/df) Pearson = 1.002679 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1656.922 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0115464 .0020885 5.53 0.000 .007453 .0156398 _cons | -1.434416 .1399661 -10.25 0.000 -1.708745 -1.160088 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1011 . 1012 . 1013 . des bf4 storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) 1014 . glm bf4 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1026.9659 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 24.75428 Deviance = 8366.946191 (1/df) Deviance = 24.75428 Pearson = 8366.946191 (1/df) Pearson = 24.75428 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.05274 Log likelihood = -1026.965868 BIC = 6396.763 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.1031788 .161925 -0.64 0.524 -.4205461 .2141884 _cons | 12.54134 .2786337 45.01 0.000 11.99523 13.08746 ------------------------------------------------------------------------------ 1015 . glm HP2sxlife bf4 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 265.0595 Iteration 2: deviance = 261.6959 Iteration 3: deviance = 261.6273 Iteration 4: deviance = 261.6271 Iteration 5: deviance = 261.6271 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 261.6270836 (1/df) Deviance = .7740446 Pearson = 301.9171445 (1/df) Pearson = .893246 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1708.557 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1431725 .0149643 -9.57 0.000 -.172502 -.1138431 _cons | .7780696 .180124 4.32 0.000 .425033 1.131106 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1016 . 1017 . des bf4m storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) 1018 . glm bf4m avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1060.6611 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 30.18079 Deviance = 10201.10756 (1/df) Deviance = 30.18079 Pearson = 10201.10756 (1/df) Pearson = 30.18079 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.250948 Log likelihood = -1060.661115 BIC = 8230.924 ------------------------------------------------------------------------------ | OIM bf4m | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.0953671 .1787945 -0.53 0.594 -.4457979 .2550637 _cons | 20.35858 .307662 66.17 0.000 19.75557 20.96159 ------------------------------------------------------------------------------ 1019 . glm HP2sxlife bf4m if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 266.1375 Iteration 2: deviance = 263.3205 Iteration 3: deviance = 263.273 Iteration 4: deviance = 263.2729 Iteration 5: deviance = 263.2729 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 263.2728985 (1/df) Deviance = .7789139 Pearson = 303.0162706 (1/df) Pearson = .8964978 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1706.911 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4m | -.1287243 .0140303 -9.17 0.000 -.1562231 -.1012255 _cons | 1.624944 .276779 5.87 0.000 1.082468 2.167421 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1020 . 1021 . des shrelaw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 1022 . glm shrelaw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1715.4166 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1420.421 Deviance = 480102.3567 (1/df) Deviance = 1420.421 Pearson = 480102.3567 (1/df) Pearson = 1420.421 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.10245 Log likelihood = -1715.416629 BIC = 478132.2 ------------------------------------------------------------------------------ | OIM shrelaw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.0649405 1.226584 -0.05 0.958 -2.469002 2.339121 _cons | 29.57199 2.110654 14.01 0.000 25.43518 33.7088 ------------------------------------------------------------------------------ 1023 . glm HP2sxlife shrelaw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 341.3096 Iteration 2: deviance = 341.0889 Iteration 3: deviance = 341.0888 Iteration 4: deviance = 341.0888 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 341.0887685 (1/df) Deviance = 1.009138 Pearson = 340.3303886 (1/df) Pearson = 1.006895 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1629.095 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shrelaw1 | .0027916 .00202 1.38 0.167 -.0011676 .0067508 _cons | -.9180819 .1010594 -9.08 0.000 -1.116155 -.7200091 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1024 . 1025 . des shfamw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 1026 . glm shfamw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1722.3986 Generalized linear models No. of obs = 339 Optimization : ML Residual df = 337 Scale parameter = 1524.9 Deviance = 513891.1861 (1/df) Deviance = 1524.9 Pearson = 513891.1861 (1/df) Pearson = 1524.9 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.17344 Log likelihood = -1722.398621 BIC = 511927.8 ------------------------------------------------------------------------------ | OIM shfamw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.8322929 1.270907 -0.65 0.513 -3.323226 1.65864 _cons | 38.72924 2.190054 17.68 0.000 34.43681 43.02167 ------------------------------------------------------------------------------ 1027 . glm HP2sxlife shfamw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 339.011 Iteration 2: deviance = 338.8015 Iteration 3: deviance = 338.8014 Iteration 4: deviance = 338.8014 Generalized linear models No. of obs = 339 Optimization : MQL Fisher scoring Residual df = 337 (IRLS EIM) Scale parameter = 1 Deviance = 338.8013645 (1/df) Deviance = 1.005345 Pearson = 339.2776273 (1/df) Pearson = 1.006759 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1624.561 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shfamw1 | .0019797 .0019787 1.00 0.317 -.0018986 .0058579 _cons | -.9179132 .1113818 -8.24 0.000 -1.136218 -.6996088 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1028 . 1029 . scalar sxlifeMedMw1 = "radhlw1" 1030 . 1031 . title4 "female impact models mediator search" ------------------------------------------------------------------------------- female impact models mediator search ------------------------------------------------------------------------------- 1032 . 1033 . * age is a mediating effect for females for Dose=> sex life for women 1034 . glm age avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 1035 . glm HP2sxlife age if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 336.5251 Iteration 2: deviance = 333.7638 Iteration 3: deviance = 333.7326 Iteration 4: deviance = 333.7326 Iteration 5: deviance = 333.7326 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 333.7325644 (1/df) Deviance = .9244669 Pearson = 379.4384762 (1/df) Pearson = 1.051076 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1794.147 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0606116 .0071416 8.49 0.000 .0466144 .0746088 _cons | -3.838422 .3937027 -9.75 0.000 -4.610065 -3.066779 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1036 . 1037 . * illness is a mediating effect for females = > sex life for men 1038 . des illw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- illw1 double %8.0g Total number of illnesses experienced in time period 1976-1986 1039 . glm illw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -259.70777 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = .2462383 Deviance = 88.89203958 (1/df) Deviance = .2462383 Pearson = 88.89203958 (1/df) Pearson = .2462383 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 1.441916 Log likelihood = -259.7077741 BIC = -2038.987 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0275426 .1585711 _cons | .1570822 .0305304 5.15 0.000 .0972436 .2169207 ------------------------------------------------------------------------------ 1040 . glm HP2sxlife illw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 412.1945 Iteration 2: deviance = 411.6441 Iteration 3: deviance = 411.6439 Iteration 4: deviance = 411.6439 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 411.6438998 (1/df) Deviance = 1.140288 Pearson = 362.6903423 (1/df) Pearson = 1.004682 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1716.236 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .257456 .1453235 1.77 0.076 -.0273728 .5422847 _cons | -.6974831 .0815339 -8.55 0.000 -.8572865 -.5376797 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1041 . 1042 . *---------- this may be important ------------------------------------------ > - 1043 . * radhlw1 can be a mediating factor for females in wave 1 for sxlife 1044 . des radhlw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- radhlw1 double %8.0g how much believed personal health is affected by radiation in 1986 1045 . glm radhlw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1821.9477 Generalized linear models No. of obs = 362 Optimization : ML Residual df = 360 Scale parameter = 1385.301 Deviance = 498708.3025 (1/df) Deviance = 1385.301 Pearson = 498708.3025 (1/df) Pearson = 1385.301 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.07706 Log likelihood = -1821.947718 BIC = 496587.3 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.789972 3.562254 1.06 0.287 -3.191917 10.77186 _cons | 55.44948 2.293757 24.17 0.000 50.9538 59.94516 ------------------------------------------------------------------------------ 1046 . glm HP2sxlife radhlw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 401.9956 Iteration 2: deviance = 401.6315 Iteration 3: deviance = 401.6315 Iteration 4: deviance = 401.6315 Generalized linear models No. of obs = 362 Optimization : MQL Fisher scoring Residual df = 360 (IRLS EIM) Scale parameter = 1 Deviance = 401.6314992 (1/df) Deviance = 1.115643 Pearson = 362.7806158 (1/df) Pearson = 1.007724 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1719.36 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0070351 .0020921 3.36 0.001 .0029347 .0111355 _cons | -1.06549 .1496137 -7.12 0.000 -1.358728 -.7722525 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1047 . *---------------------------------------------------------------------------- > - 1048 . 1049 . des bf4 // soma recentered storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) 1050 . * bf4 is a mediting effect for females for Dose=> sex life for women 1051 . glm bf4 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 1052 . glm HP2sxlife bf4 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 344.0399 Iteration 2: deviance = 342.6827 Iteration 3: deviance = 342.6686 Iteration 4: deviance = 342.6686 Iteration 5: deviance = 342.6686 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 342.6686152 (1/df) Deviance = .9492205 Pearson = 336.3540555 (1/df) Pearson = .9317287 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1785.211 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1230087 .0148291 -8.30 0.000 -.1520732 -.0939442 _cons | .5134392 .1542351 3.33 0.001 .211144 .8157344 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1053 . 1054 . des bf4m // soma recentered storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) 1055 . * bf4m is a possible mediating effect for female sex life 1056 . glm bf4m avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1140.521 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 31.548 Deviance = 11388.82943 (1/df) Deviance = 31.548 Pearson = 11388.82943 (1/df) Pearson = 31.548 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.294882 Log likelihood = -1140.521046 BIC = 9260.95 ------------------------------------------------------------------------------ | OIM bf4m | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.563541 .5374133 -2.91 0.004 -2.616852 -.5102303 _cons | 18.82212 .3455741 54.47 0.000 18.1448 19.49943 ------------------------------------------------------------------------------ 1057 . glm HP2sxlife bf4m if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 356.1451 Iteration 2: deviance = 355.6209 Iteration 3: deviance = 355.6178 Iteration 4: deviance = 355.6178 Iteration 5: deviance = 355.6178 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 355.6178202 (1/df) Deviance = .9850909 Pearson = 340.0032816 (1/df) Pearson = .9418373 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1772.262 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4m | -.0990193 .0135516 -7.31 0.000 -.12558 -.0724586 _cons | 1.063679 .244399 4.35 0.000 .584666 1.542693 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1058 . 1059 . des shfamw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 1060 . glm shfamw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1820.2144 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 1334.587 Deviance = 481786.0787 (1/df) Deviance = 1334.587 Pearson = 481786.0787 (1/df) Pearson = 1334.587 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.03975 Log likelihood = -1820.214442 BIC = 479658.2 ------------------------------------------------------------------------------ | OIM shfamw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 2.325224 3.495393 0.67 0.506 -4.525621 9.176069 _cons | 34.91682 2.24765 15.53 0.000 30.5115 39.32213 ------------------------------------------------------------------------------ 1061 . glm HP2sxlife shfamw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 415.7461 Iteration 2: deviance = 415.0916 Iteration 3: deviance = 415.0914 Iteration 4: deviance = 415.0914 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 415.0913862 (1/df) Deviance = 1.149838 Pearson = 362.9739265 (1/df) Pearson = 1.005468 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1712.788 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shfamw1 | .0007609 .0020813 0.37 0.715 -.0033182 .0048401 _cons | -.6739948 .1071104 -6.29 0.000 -.8839274 -.4640622 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1062 . 1063 . des shrelaw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 1064 . glm shrelaw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1801.2674 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 1202.293 Deviance = 434027.9437 (1/df) Deviance = 1202.293 Pearson = 434027.9437 (1/df) Pearson = 1202.293 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 9.935358 Log likelihood = -1801.267425 BIC = 431900.1 ------------------------------------------------------------------------------ | OIM shrelaw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 5.240218 3.317628 1.58 0.114 -1.262215 11.74265 _cons | 26.01592 2.133342 12.19 0.000 21.83464 30.19719 ------------------------------------------------------------------------------ 1065 . glm HP2sxlife shrelaw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 414.3957 Iteration 2: deviance = 413.7847 Iteration 3: deviance = 413.7844 Iteration 4: deviance = 413.7844 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 413.7844342 (1/df) Deviance = 1.146217 Pearson = 362.9072137 (1/df) Pearson = 1.005283 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1714.095 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shrelaw1 | .002446 .0021561 1.13 0.257 -.0017799 .0066718 _cons | -.7168064 .0987075 -7.26 0.000 -.9102696 -.5233433 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1066 . 1067 . glm aborw1 avgcumdosew1 if gender==2, fam(pois) link(log) Iteration 0: log likelihood = -294.95486 Iteration 1: log likelihood = -274.67418 Iteration 2: log likelihood = -274.64155 Iteration 3: log likelihood = -274.64155 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 1 Deviance = 418.9504185 (1/df) Deviance = 1.160527 Pearson = 1079.525061 (1/df) Pearson = 2.990374 Variance function: V(u) = u [Poisson] Link function : g(u) = ln(u) [Log] AIC = 1.524196 Log likelihood = -274.6415467 BIC = -1708.929 ------------------------------------------------------------------------------ | OIM aborw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0974558 .154214 0.63 0.527 -.204798 .3997097 _cons | -1.333588 .1156098 -11.54 0.000 -1.560179 -1.106997 ------------------------------------------------------------------------------ 1068 . glm HP2sxlife aborw1 if gender==2, fam(bin) link(probit) irls scale(dev) Iteration 1: deviance = 414.3147 Iteration 2: deviance = 413.5468 Iteration 3: deviance = 413.545 Iteration 4: deviance = 413.545 Iteration 5: deviance = 413.545 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 413.5449635 (1/df) Deviance = 1.145554 Pearson = 362.1757599 (1/df) Pearson = 1.003257 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1714.334 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- aborw1 | -.1290317 .1159726 -1.11 0.266 -.3563338 .0982703 _cons | -.615959 .0802001 -7.68 0.000 -.7731484 -.4587696 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1069 . 1070 . title4 "h1 pt2 wave 1 sex life impact summary matrix construction" ------------------------------------------------------------------------------- h1 pt2 wave 1 sex life impact summary matrix construction ------------------------------------------------------------------------------- 1071 . 1072 . *xx summary of mediating effects: age and illness mediate sex life for men 1073 . * age illnesss radhlw1 bf4 bf4m (soma) media > te sex life for women 1074 . 1075 . scalar SigDoseMEhmcrW1 = "no" 1076 . scalar MaineffhmcrMw1 = "age bf4 bf40" 1077 . 1078 . scalar MainEffsxlifeFw1 = "age radhlw1 bf4 bf4m" 1079 . scalar SigDoseSxlifeFw1="no" 1080 . scalar SxLifeModFw1 = "no" 1081 . scalar sxlifeModMw1 = "none" 1082 . * xx female main effects modelscalar SigDoseSxlifeFw1 = "no" 1083 . scalar MainEffsxlifeFw1 = "age bf4 bf4m" 1084 . scalar sxlifeMedMw1 = "age illw1" 1085 . scalar sxlifeMedFw1 = "age illw1 radhlw1 bf4 bf4m" 1086 . scalar sxlifeMedMw1 = "radhlw1" 1087 . 1088 . *--- summary matrix contstruction 1089 . 1090 . matrix define sxlifeMw1 = J(1,8, 0) 1091 . matrix define sxlifeFw1 = J(1,8, 0) 1092 . matrix colnames sxlifeMw1= hypnum ptnum wave gender medsig numMAsig numMods > ig numMed 1093 . matrix colnames sxlifeFw1= hypnum ptnum wave gender medsig numMAsig > numModsig numMed 1094 . matrix define sxlifeMw1= (1, 2, 1, 1, 0, 3, 0, 1 ) 1095 . matrix define sxlifeFw1= (1, 2, 1, 2, 0, 4, 0, 5) 1096 . matrix rowname sxlifeMw1 = sxlifeMw1 1097 . matrix rowname sxlifeFw1 = sxlifeFw1 1098 . matlist sxlifeMw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- sxlifeMw1 | 1 2 1 1 0 3 > 0 1 1099 . matlist sxlifeFw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- sxlifeFw1 | 1 2 1 2 0 4 > 0 5 1100 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1 \ hmcrMw1 \ hmcrFw1 \ spM > w1 \ spFw1 \ prbfamMw1 \ prbfamFw1 \ sxlifeMw1 \ sxlifeFw1 ) 1101 . 1102 . matlist H1pt2w1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 1 > 1 2 r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 2 > 0 2 spMw1 | 1 2 1 1 1 2 > 2 1 spFw1 | 1 2 1 2 0 5 > 0 2 prbfamMw1 | 1 2 1 1 0 2 > 0 0 prbfamFw1 | 1 2 1 2 0 3 > 0 2 sxlifeMw1 | 1 2 1 1 0 3 > 0 1 sxlifeFw1 | 1 2 1 2 0 4 > 0 5 1103 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 1104 . matrix rownames H1pt2w1 = wkMw1 wkFw1 hmcrMw1 hmcrFw1 spMw1 spFw1 pr > bfhmMw1 prbfhmFw1 1105 . matlist H1pt2w1 | hypnum ptnum wave gender medsig numMAsig > numModsig numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 hmcrMw1 | 1 2 1 1 0 2 > 0 1 hmcrFw1 | 1 2 1 2 0 2 > 0 2 spMw1 | 1 2 1 1 1 2 > 2 1 spFw1 | 1 2 1 2 0 5 > 0 2 prbfhmMw1 | 1 2 1 1 0 2 > 0 0 prbfhmFw1 | 1 2 1 2 0 3 > 0 2 prbfhmFw1 | 1 2 1 1 0 3 > 0 1 prbfhmFw1 | 1 2 1 2 0 4 > 0 5 1106 . 1107 . 1108 . 1109 . 1110 . *=================== Chunk 8 Dose => interests and Hobbies relationship 1111 . 1112 . title " 6. H1 wave 1 part2 Dose-Interest and Hobbies impact" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 6. H1 wave 1 part2 Dose-Interest and Hobbies impact ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:17:37 ***** ******************************************************************************* ******************************************************************************* 1113 . 1114 . 1115 . * Chunk 8 ---male models 1116 . forvalues j=1/1 { 2. set more off 3. 1117 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 1118 . foreach var in HP2inthob { 5. forvalues k=1/2 { 6. local w1bf bf1 bf4 bf9 bf10 bf11 bf14 bf15m bf20 bf22 bf30 bf40 7. 1119 . di as input "Full main model for `var' for wave= `j' " 8. di _skip(4) 9. di as input "chunk 8 H1 test:Gender= `k' model Wave = `j' for `e(depvar > )' " 10. di _skip(4) 11. title "Full Nottingham Part 2 subscale models for male and then females" 12. di as input "Model for gender==`k' and wave == `j'" 13. di _skip(2) 14. xi: logistic `var' age i.educ occ1w`j'-occ8w`j' /// > marrw`j'1- marrw`j'3 marrw`j'5-marrw`j'6 inc1w`j'-inc4w`j' // > / > radhlw`j' havmil avgcumdosew`j' `w`j'bf' /// > deaw`j' dvcew`j' sepaw`j' accdw`j' movew`j' /// > illw`j' shfamw`j' shhlw`j' shjobw`j' shrelaw`j' suprtw`j' suc > hrw`j' /// > havmilsq radhlw1 if gender==`k', coef difficult iterate(50) > 15. estat class 16. estat gof 17. fitstat 18. } 19. } 20. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2inthob for wave= 1 chunk 8 H1 test:Gender= 1 model Wave = 1 for HP2sxlife ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Full Nottingham Part 2 subscale models for male and then females ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:17:37 ***** ******************************************************************************* ******************************************************************************* Model for gender==1 and wave == 1 i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: _Ieduc_4 != 0 predicts failure perfectly _Ieduc_4 dropped and 12 obs not used note: _Ieduc_7 != 0 predicts failure perfectly _Ieduc_7 dropped and 4 obs not used note: _Ieduc_8 != 0 predicts failure perfectly _Ieduc_8 dropped and 2 obs not used note: occ6w1 != 0 predicts failure perfectly occ6w1 dropped and 4 obs not used note: occ7w1 != 0 predicts failure perfectly occ7w1 dropped and 4 obs not used note: marrw15 != 0 predicts success perfectly marrw15 dropped and 1 obs not used note: bf15m != 0 predicts failure perfectly bf15m dropped and 19 obs not used note: dvcew1 != 0 predicts failure perfectly dvcew1 dropped and 2 obs not used note: sepaw1 != 0 predicts failure perfectly sepaw1 dropped and 1 obs not used note: _Ieduc_6 omitted because of collinearity note: marrw16 omitted because of collinearity note: radhlw1 omitted because of collinearity convergence not achieved Logistic regression Number of obs = 284 LR chi2(40) = 100.41 Prob > chi2 = 0.0000 Log likelihood = -55.819677 Pseudo R2 = 0.4735 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1670729 .0496836 3.36 0.001 .0696948 .2644509 _Ieduc_2 | -3.680581 2.320262 -1.59 0.113 -8.228211 .8670494 _Ieduc_3 | -.4101129 .8396848 -0.49 0.625 -2.055865 1.235639 _Ieduc_4 | 0 (omitted) _Ieduc_5 | -.2315388 .8579156 -0.27 0.787 -1.913023 1.449945 _Ieduc_6 | 0 (omitted) _Ieduc_7 | 0 (omitted) _Ieduc_8 | 0 (omitted) occ1w1 | -.855187 2.512009 -0.34 0.734 -5.778634 4.06826 occ2w1 | -.8206155 2.564147 -0.32 0.749 -5.846251 4.20502 occ3w1 | .4752119 2.731393 0.17 0.862 -4.87822 5.828644 occ4w1 | -.0629245 2.648929 -0.02 0.981 -5.254729 5.12888 occ5w1 | .1433339 2.754114 0.05 0.958 -5.254631 5.541299 occ6w1 | 0 (omitted) occ7w1 | 0 (omitted) occ8w1 | 2.572986 2.905992 0.89 0.376 -3.122655 8.268626 marrw11 | 13.32407 13699.24 0.00 0.999 -26836.69 26863.34 marrw12 | 14.0947 13699.24 0.00 0.999 -26835.93 26864.11 marrw13 | 12.02184 13699.24 0.00 0.999 -26838 26862.04 marrw15 | 0 (omitted) marrw16 | 0 (omitted) inc1w1 | -2.316757 2.703555 -0.86 0.391 -7.615628 2.982114 inc2w1 | -1.011499 2.582483 -0.39 0.695 -6.073072 4.050074 inc3w1 | -1.18772 2.503558 -0.47 0.635 -6.094604 3.719164 inc4w1 | -3.273623 3.286978 -1.00 0.319 -9.715981 3.168734 radhlw1 | .0108481 .0098238 1.10 0.269 -.0084061 .0301023 havmil | .011552 .0121868 0.95 0.343 -.0123338 .0354378 avgcumdosew1 | -.0246271 .287731 -0.09 0.932 -.5885695 .5393153 bf1 | -1.705805 342.4811 -0.00 0.996 -672.9563 669.5447 bf4 | -.2162507 .0773979 -2.79 0.005 -.3679478 -.0645536 bf9 | -.0313278 .0435662 -0.72 0.472 -.116716 .0540605 bf10 | -.058867 .0329385 -1.79 0.074 -.1234253 .0056913 bf11 | -.0066494 .0862166 -0.08 0.939 -.1756309 .162332 bf14 | -.0001753 .0001131 -1.55 0.121 -.0003971 .0000464 bf15m | 0 (omitted) bf20 | 1.703694 342.481 0.00 0.996 -669.5468 672.9542 bf22 | .0001385 .0001388 1.00 0.318 -.0001335 .0004105 bf30 | .0005239 .0005317 0.99 0.324 -.0005182 .0015661 bf40 | .0218384 .248862 0.09 0.930 -.4659222 .509599 deaw1 | -1.453389 1.021447 -1.42 0.155 -3.455388 .5486112 dvcew1 | 0 (omitted) sepaw1 | 0 (omitted) accdw1 | -.0050761 1.624722 -0.00 0.998 -3.189472 3.17932 movew1 | -.0513738 1.125065 -0.05 0.964 -2.256461 2.153713 illw1 | -.5020103 .7607393 -0.66 0.509 -1.993032 .9890113 shfamw1 | .0137784 .0108362 1.27 0.204 -.0074601 .0350169 shhlw1 | .0112349 .0171506 0.66 0.512 -.0223798 .0448495 shjobw1 | -.0014008 .0136329 -0.10 0.918 -.0281207 .0253192 shrelaw1 | -.0197593 .0117959 -1.68 0.094 -.0428787 .0033602 suprtw1 | .0328165 .033418 0.98 0.326 -.0326817 .0983146 suchrw1 | .0158253 .0298163 0.53 0.596 -.0426135 .0742642 havmilsq | -.000019 .0000262 -0.72 0.469 -.0000704 .0000324 radhlw1 | 0 (omitted) _cons | -89.90699 . . . . . ------------------------------------------------------------------------------ Note: 29 failures and 0 successes completely determined. Warning: convergence not achieved Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 19 7 | 26 - | 16 242 | 258 -----------+--------------------------+----------- Total | 35 249 | 284 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 54.29% Specificity Pr( -|~D) 97.19% Positive predictive value Pr( D| +) 73.08% Negative predictive value Pr(~D| -) 93.80% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.81% False - rate for true D Pr( -| D) 45.71% False + rate for classified + Pr(~D| +) 26.92% False - rate for classified - Pr( D| -) 6.20% -------------------------------------------------- Correctly classified 91.90% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 284 number of covariate patterns = 284 Pearson chi2(242) = 148.05 Prob > chi2 = 1.0000 Measures of Fit for logistic of HP2inthob Log-Lik Intercept Only: -106.026 Log-Lik Full Model: -55.820 D(230): 111.639 LR(40): 100.412 Prob > LR: 0.000 McFadden's R2: 0.474 McFadden's Adj R2: -0.036 Maximum Likelihood R2: 0.298 Cragg & Uhler's R2: 0.566 McKelvey and Zavoina's R2: 0.975 Efron's R2: 0.438 Variance of y*: 130.454 Variance of error: 3.290 Count R2: 0.919 Adj Count R2: 0.343 AIC: 0.773 AIC*n: 219.639 BIC: -1187.625 BIC': 125.547 Full main model for HP2inthob for wave= 1 chunk 8 H1 test:Gender= 2 model Wave = 1 for HP2inthob ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Full Nottingham Part 2 subscale models for male and then females ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:17:39 ***** ******************************************************************************* ******************************************************************************* Model for gender==2 and wave == 1 i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: bf15m != 0 predicts failure perfectly bf15m dropped and 11 obs not used note: _Ieduc_8 omitted because of collinearity note: radhlw1 omitted because of collinearity Logistic regression Number of obs = 347 LR chi2(50) = 136.35 Prob > chi2 = 0.0000 Log likelihood = -100.6444 Pseudo R2 = 0.4038 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1576323 .0314369 5.01 0.000 .0960172 .2192474 _Ieduc_2 | -13.15998 1390.096 -0.01 0.992 -2737.698 2711.378 _Ieduc_3 | -12.34293 1390.096 -0.01 0.993 -2736.881 2712.195 _Ieduc_4 | -11.76283 1390.096 -0.01 0.993 -2736.302 2712.776 _Ieduc_5 | -12.38126 1390.096 -0.01 0.993 -2736.92 2712.157 _Ieduc_6 | -12.87444 1390.096 -0.01 0.993 -2737.413 2711.664 _Ieduc_7 | -11.28064 1390.098 -0.01 0.994 -2735.823 2713.262 _Ieduc_8 | 0 (omitted) occ1w1 | -2.007966 2.670409 -0.75 0.452 -7.241872 3.225941 occ2w1 | -1.42005 2.705 -0.52 0.600 -6.721752 3.881652 occ3w1 | -2.637816 2.721093 -0.97 0.332 -7.971061 2.695429 occ4w1 | -1.022244 2.749637 -0.37 0.710 -6.411434 4.366946 occ5w1 | -2.685007 2.926173 -0.92 0.359 -8.4202 3.050187 occ6w1 | -1.851042 2.900638 -0.64 0.523 -7.536188 3.834105 occ7w1 | -.7751235 2.748828 -0.28 0.778 -6.162727 4.61248 occ8w1 | 1.738342 2.770527 0.63 0.530 -3.691792 7.168476 marrw11 | -1.712685 1.961435 -0.87 0.383 -5.557026 2.131656 marrw12 | -.7978205 2.432804 -0.33 0.743 -5.566028 3.970387 marrw13 | -2.20107 1.785855 -1.23 0.218 -5.701282 1.299141 marrw15 | -2.131202 2.204296 -0.97 0.334 -6.451543 2.189139 marrw16 | -2.419024 2.291199 -1.06 0.291 -6.909692 2.071645 inc1w1 | .4393464 2.638625 0.17 0.868 -4.732264 5.610957 inc2w1 | 1.095503 2.586215 0.42 0.672 -3.973385 6.16439 inc3w1 | 1.162288 2.607351 0.45 0.656 -3.948028 6.272603 inc4w1 | 2.352437 2.762052 0.85 0.394 -3.061086 7.765959 radhlw1 | .016256 .0073318 2.22 0.027 .001886 .030626 havmil | .0027093 .0039697 0.68 0.495 -.0050712 .0104899 avgcumdosew1 | .4525599 .2626453 1.72 0.085 -.0622155 .9673353 bf1 | -.0392934 .0467729 -0.84 0.401 -.1309666 .0523799 bf4 | -.1618817 .055024 -2.94 0.003 -.2697268 -.0540365 bf9 | -.0557262 .0332131 -1.68 0.093 -.1208228 .0093703 bf10 | .0039258 .0156035 0.25 0.801 -.0266565 .0345082 bf11 | .0160165 .0531713 0.30 0.763 -.0881973 .1202303 bf14 | -.00005 .0000899 -0.56 0.578 -.0002263 .0001262 bf15m | 0 (omitted) bf20 | .0307621 .0426341 0.72 0.471 -.0527993 .1143235 bf22 | .0000428 .0000948 0.45 0.651 -.000143 .0002287 bf30 | .0009017 .0003472 2.60 0.009 .0002212 .0015823 bf40 | -.134405 .1339765 -1.00 0.316 -.396994 .1281841 deaw1 | .0546666 .2268628 0.24 0.810 -.3899763 .4993096 dvcew1 | 1.590367 2.621038 0.61 0.544 -3.546772 6.727507 sepaw1 | -1.421464 2.957733 -0.48 0.631 -7.218515 4.375587 accdw1 | -.187097 1.01096 -0.19 0.853 -2.168543 1.794349 movew1 | .280486 .7016446 0.40 0.689 -1.094712 1.655684 illw1 | .0230342 .4319012 0.05 0.957 -.8234766 .8695451 shfamw1 | .0000591 .0083223 0.01 0.994 -.0162524 .0163705 shhlw1 | -.0037244 .0125633 -0.30 0.767 -.0283479 .0208992 shjobw1 | -.0032054 .0093218 -0.34 0.731 -.0214757 .0150649 shrelaw1 | -.0051493 .0079574 -0.65 0.518 -.0207454 .0104468 suprtw1 | .0012475 .0121595 0.10 0.918 -.0225847 .0250796 suchrw1 | -.0019168 .0162004 -0.12 0.906 -.0336689 .0298353 havmilsq | -1.74e-06 5.49e-06 -0.32 0.752 -.0000125 9.02e-06 radhlw1 | 0 (omitted) _cons | 3.925606 1390.1 0.00 0.998 -2720.621 2728.472 ------------------------------------------------------------------------------ Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 37 7 | 44 - | 29 274 | 303 -----------+--------------------------+----------- Total | 66 281 | 347 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 56.06% Specificity Pr( -|~D) 97.51% Positive predictive value Pr( D| +) 84.09% Negative predictive value Pr(~D| -) 90.43% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 2.49% False - rate for true D Pr( -| D) 43.94% False + rate for classified + Pr(~D| +) 15.91% False - rate for classified - Pr( D| -) 9.57% -------------------------------------------------- Correctly classified 89.63% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 347 number of covariate patterns = 347 Pearson chi2(296) = 450.83 Prob > chi2 = 0.0000 Measures of Fit for logistic of HP2inthob Log-Lik Intercept Only: -168.821 Log-Lik Full Model: -100.644 D(293): 201.289 LR(50): 136.353 Prob > LR: 0.000 McFadden's R2: 0.404 McFadden's Adj R2: 0.084 Maximum Likelihood R2: 0.325 Cragg & Uhler's R2: 0.522 McKelvey and Zavoina's R2: 0.648 Efron's R2: 0.466 Variance of y*: 9.347 Variance of error: 3.290 Count R2: 0.896 Adj Count R2: 0.455 AIC: 0.891 AIC*n: 309.289 BIC: -1512.563 BIC': 156.113 1120 . 1121 . label var radhlw1 "Self-perceived Chornobyl health threat in wave 1" 1122 . 1123 . title4 "7. Testing H1 Wv2 Moderators of male Dose => Interests and Hobbies Im > pact" ------------------------------------------------------------------------------- 7. Testing H1 Wv2 Moderators of male Dose => Interests and Hobbies Impact ------------------------------------------------------------------------------- 1124 . 1125 . 1126 . forvalues j=1/1 { 2. set more off 3. 1127 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf4 bf14 bf40 4. 1128 . foreach var in HP2inthob { 5. local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 6. di _skip(2) 7. di as input "Full main model for `var' for wave= `j' " 8. di _skip(4) 9. di as input "chunk 8 H1 test:Gender= male model Wave = `j' for `e(depva > r)' " 10. di _skip(4) 11. title "Full Nottingham Part 2 subscale models " "males on wave=`j'" 12. des bf5m shfamw1 radhlw1 bf4m 13. xi: logistic `var' age radhlw1 bf4 bf10 bf40 shrelaw1 /// > suchrw1 if gender==1, coef difficult iterate(50) 14. estat class 15. estat gof 16. fitstat 17. di _skip(2) 18. di as input "Note: bf4m is necssary for bf5 but if bf4m is in model bf5 is > not signif." 19. di as input " Therefore, bf5 is not deemed significant." 20. } 21. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf14 float %9.0g bf14= max(0, radw2 - 10) * bf12 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2inthob for wave= 1 chunk 8 H1 test:Gender= male model Wave = 1 for HP2inthob ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Full Nottingham Part 2 subscale models ***** ***** males on wave=1 ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:17:40 ***** ******************************************************************************* ******************************************************************************* storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf5m float %9.0g bf5m = max(0, ecprw3 - 75) * bf4m shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 radhlw1 double %8.0g Self-perceived Chornobyl health threat in wave 1 bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) Logistic regression Number of obs = 340 LR chi2(7) = 64.52 Prob > chi2 = 0.0000 Log likelihood = -86.806728 Pseudo R2 = 0.2709 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0200969 .0187425 1.07 0.284 -.0166377 .0568315 radhlw1 | .0186628 .006222 3.00 0.003 .0064679 .0308576 bf4 | -.1419329 .0385628 -3.68 0.000 -.2175146 -.0663512 bf10 | -.0002913 .0097452 -0.03 0.976 -.0193915 .0188088 bf40 | .2956738 .1272383 2.32 0.020 .0462913 .5450564 shrelaw1 | -.0046332 .0053472 -0.87 0.386 -.0151134 .005847 suchrw1 | .0141811 .015058 0.94 0.346 -.015332 .0436942 _cons | -3.536445 1.260759 -2.81 0.005 -6.007486 -1.065403 ------------------------------------------------------------------------------ Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 13 4 | 17 - | 25 298 | 323 -----------+--------------------------+----------- Total | 38 302 | 340 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 34.21% Specificity Pr( -|~D) 98.68% Positive predictive value Pr( D| +) 76.47% Negative predictive value Pr(~D| -) 92.26% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.32% False - rate for true D Pr( -| D) 65.79% False + rate for classified + Pr(~D| +) 23.53% False - rate for classified - Pr( D| -) 7.74% -------------------------------------------------- Correctly classified 91.47% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 340 number of covariate patterns = 332 Pearson chi2(324) = 299.10 Prob > chi2 = 0.8360 Measures of Fit for logistic of HP2inthob Log-Lik Intercept Only: -119.064 Log-Lik Full Model: -86.807 D(332): 173.613 LR(7): 64.515 Prob > LR: 0.000 McFadden's R2: 0.271 McFadden's Adj R2: 0.204 Maximum Likelihood R2: 0.173 Cragg & Uhler's R2: 0.343 McKelvey and Zavoina's R2: 0.425 Efron's R2: 0.237 Variance of y*: 5.726 Variance of error: 3.290 Count R2: 0.915 Adj Count R2: 0.237 AIC: 0.558 AIC*n: 189.613 BIC: -1761.596 BIC': -23.712 Note: bf4m is necssary for bf5 but if bf4m is in model bf5 is not signif. Therefore, bf5 is not deemed significant. 1129 . 1130 . cap gen bf4Xd1 = bf4*avgcumdosew1 1131 . cap gen bf10Xd1= bf10*avgcumdosew1 1132 . cap gen bf40Xd1 = bf40*agecumdosew1 1133 . 1134 . scalar SigdoseMEinthob = "no" 1135 . scalar MainEffMw1 = "radhlw1 bf4 bf40" 1136 . 1137 . title3 "wave 1 Main effects Dose=> Interests and Hobbies impact identificatio > n" ------------------------------------------------------------------------------- title3 : wave 1 Main effects Dose=> Interests and Hobbies impact identification 18 Jun 2012 18:17:42 computer Macintosh (Intel 64-bit) folder /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/ > h1pt2 Data file chwide16june2012.dta currrently has 2380 variables and 703 obse > rvations 1138 . forvalues j=1/1 { 2. set more off 3. 1139 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 1140 . foreach var in HP2inthob { 5. local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 6. di as input "Full main model for `var' for wave= `j' " 7. di _skip(4) 8. di as input "chunk 8 H1 test:Gender= male model Wave = `j' for `e(depva > r)' " 9. di _skip(4) 10. 1141 . 1142 . xi: logistic `var' age /// > radhlw`j' avgcumdosew`j' /// > shfamw`j' bf4 /// > bf40 bf4Xd1 bf40Xd1 if gender==1, coef difficult iterate(50 > ) 11. estat class 12. estat gof 13. fitstat 14. } 15. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2inthob for wave= 1 chunk 8 H1 test:Gender= male model Wave = 1 for HP2inthob Logistic regression Number of obs = 339 LR chi2(8) = 69.68 Prob > chi2 = 0.0000 Log likelihood = -84.103355 Pseudo R2 = 0.2929 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0229294 .0193857 1.18 0.237 -.015066 .0609247 radhlw1 | .0203775 .0063075 3.23 0.001 .008015 .03274 avgcumdosew1 | 6.417441 3.545095 1.81 0.070 -.530818 13.3657 shfamw1 | .0031998 .0052579 0.61 0.543 -.0071055 .013505 bf4 | .0015257 .0840489 0.02 0.986 -.163207 .1662585 bf40 | .4371492 .1715289 2.55 0.011 .1009587 .7733397 bf4Xd1 | -.469083 .2679615 -1.75 0.080 -.9942779 .0561118 bf40Xd1 | -.724812 .4037974 -1.79 0.073 -1.51624 .0666164 _cons | -5.810547 1.708506 -3.40 0.001 -9.159156 -2.461938 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 11 6 | 17 - | 27 295 | 322 -----------+--------------------------+----------- Total | 38 301 | 339 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 28.95% Specificity Pr( -|~D) 98.01% Positive predictive value Pr( D| +) 64.71% Negative predictive value Pr(~D| -) 91.61% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 1.99% False - rate for true D Pr( -| D) 71.05% False + rate for classified + Pr(~D| +) 35.29% False - rate for classified - Pr( D| -) 8.39% -------------------------------------------------- Correctly classified 90.27% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 339 number of covariate patterns = 337 Pearson chi2(328) = 260.58 Prob > chi2 = 0.9975 Measures of Fit for logistic of HP2inthob Log-Lik Intercept Only: -118.946 Log-Lik Full Model: -84.103 D(330): 168.207 LR(8): 69.684 Prob > LR: 0.000 McFadden's R2: 0.293 McFadden's Adj R2: 0.217 Maximum Likelihood R2: 0.186 Cragg & Uhler's R2: 0.368 McKelvey and Zavoina's R2: 0.640 Efron's R2: 0.242 Variance of y*: 9.132 Variance of error: 3.290 Count R2: 0.903 Adj Count R2: 0.132 AIC: 0.549 AIC*n: 186.207 BIC: -1754.373 BIC': -23.076 1143 . scalar SigDoseInthbMw1 = "no" 1144 . scalar MainEffInthbMw1 = "age radhlw1 shfamw1" 1145 . scalar InthbModMw1 = "none" 1146 . 1147 . *------chunk 8 female moderator models 1148 . title4 "trimmed Moderators of female Dose => Interests and Hobbies Impact" ------------------------------------------------------------------------------- trimmed Moderators of female Dose => Interests and Hobbies Impact ------------------------------------------------------------------------------- 1149 . 1150 . 1151 . forvalues j=1/1 { 2. set more off 3. 1152 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 1153 . foreach var in HP2inthob { 5. local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 6. di as input "Full main model for `var' for wave= `j' " 7. di _skip(4) 8. di as input "chunk 8 H1 test:Gender= male model Wave = `j' for `e(depva > r)' " 9. di _skip(4) 10. title "Full Nottingham Part 2 subscale models for females " 11. 1154 . xi: logistic `var' age /// > radhlw`j' avgcumdosew`j' /// > bf4 /// > if gender==2, coef difficult iterate(50) 12. estat class 13. estat gof 14. fitstat 15. } 16. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2inthob for wave= 1 chunk 8 H1 test:Gender= male model Wave = 1 for HP2inthob ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Full Nottingham Part 2 subscale models for females ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:17:43 ***** ******************************************************************************* ******************************************************************************* Logistic regression Number of obs = 362 LR chi2(4) = 91.22 Prob > chi2 = 0.0000 Log likelihood = -126.30289 Pseudo R2 = 0.2653 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0702541 .0161129 4.36 0.000 .0386734 .1018347 radhlw1 | .0195656 .004896 4.00 0.000 .0099695 .0291616 avgcumdosew1 | .2200581 .222466 0.99 0.323 -.2159672 .6560835 bf4 | -.1103614 .0315756 -3.50 0.000 -.1722484 -.0484743 _cons | -5.707394 1.074147 -5.31 0.000 -7.812684 -3.602104 ------------------------------------------------------------------------------ Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 29 10 | 39 - | 37 286 | 323 -----------+--------------------------+----------- Total | 66 296 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 43.94% Specificity Pr( -|~D) 96.62% Positive predictive value Pr( D| +) 74.36% Negative predictive value Pr(~D| -) 88.54% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 3.38% False - rate for true D Pr( -| D) 56.06% False + rate for classified + Pr(~D| +) 25.64% False - rate for classified - Pr( D| -) 11.46% -------------------------------------------------- Correctly classified 87.02% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 362 number of covariate patterns = 360 Pearson chi2(355) = 441.04 Prob > chi2 = 0.0012 Measures of Fit for logistic of HP2inthob Log-Lik Intercept Only: -171.912 Log-Lik Full Model: -126.303 D(357): 252.606 LR(4): 91.217 Prob > LR: 0.000 McFadden's R2: 0.265 McFadden's Adj R2: 0.236 Maximum Likelihood R2: 0.223 Cragg & Uhler's R2: 0.363 McKelvey and Zavoina's R2: 0.426 Efron's R2: 0.309 Variance of y*: 5.731 Variance of error: 3.290 Count R2: 0.870 Adj Count R2: 0.288 AIC: 0.725 AIC*n: 262.606 BIC: -1850.711 BIC': -67.651 1155 . scalar SigdoseInthbFw1 = "no" 1156 . scalar MainEffInthbFw1 = "age radhlw1 bf4" 1157 . 1158 . title4 "*------chunk 8 testing female interests and hobbies moderators" ------------------------------------------------------------------------------- *------chunk 8 testing female interests and hobbies moderators ------------------------------------------------------------------------------- 1159 . 1160 . 1161 . forvalues j=1/1 { 2. set more off 3. 1162 . des age educ1-educ7 marrw`j'1-marrw`j'6 inc1w`j'-inc4w`j' /// > bf1 bf4 bf9 bf11 bf4m bf15m bf30 bf40 4. 1163 . cap gen ageXd1 = age*avgcumdosew1 5. foreach var in HP2inthob { 6. local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 7. di _skip(4) 8. di as input "Full main model for `var' for wave= `j' " 9. di _skip(4) 10. di as input "chunk 8 H1 test:Gender= male model Wave = `j' for `e(depva > r)' " 11. di _skip(4) 12. 1164 . xi: logistic `var' age /// > radhlw`j' avgcumdosew`j' /// > bf4 bf4Xd1 ageXd1 radhlw1Xd1 /// > if gender==2, coef difficult iterate(50) 13. estat class 14. estat gof 15. fitstat 16. } 17. } storage display value variable name type format label variable label ------------------------------------------------------------------------------- age double %8.0g * Respondent's age educ1 byte %8.0g educ==1. did not graduate high school educ2 byte %8.0g educ==2. graduated high school educ3 byte %8.0g educ==3. technical degree educ4 byte %8.0g educ==4. did not finish college/bachelor's educ5 byte %8.0g educ==5. graduated college/bachelor's educ6 byte %8.0g educ==6. finished specialist/master's degree educ7 byte %8.0g educ==7. doctor of science/phd marrw11 byte %8.0g marrw1==1. single marrw12 byte %8.0g marrw1==2. cohabitating marrw13 byte %8.0g marrw1==3. married marrw14 byte %8.0g marrw1==4. separated marrw15 byte %8.0g marrw1==5. divorced marrw16 byte %8.0g marrw1==6. widowed inc1w1 double %15.0g LABJ Income is not sufficient for basic neccessities in 1986 inc2w1 double %15.0g LABJ Income is just sufficient for basic neccessities in 1986 inc3w1 double %15.0g LABJ Income is sufficient for basics plus extra purchases/savings in 1986 inc4w1 double %15.0g LABJ Income allows to comfortably afford luxury items in 1986 bf1 float %9.0g bf1 = max(0, kzchorn - 40) bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) bf9 float %9.0g bf9= max(0, 30 - shhlw1) bf11 float %9.0g bf11= max(0, 20 - sufamw1) bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) bf15m float %9.0g bf15m= max(0, 1 - icdxcnt) * bf2 bf30 float %9.0g bf30 = max(0, neiw1 - 85) * bf20 bf40 float %9.0g bf40 = max(0, icdxcnt - 1.01635E-007) Full main model for HP2inthob for wave= 1 chunk 8 H1 test:Gender= male model Wave = 1 for HP2inthob Logistic regression Number of obs = 362 LR chi2(7) = 91.60 Prob > chi2 = 0.0000 Log likelihood = -126.114 Pseudo R2 = 0.2664 ------------------------------------------------------------------------------ HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .067857 .0189391 3.58 0.000 .030737 .1049771 radhlw1 | .0207366 .0056517 3.67 0.000 .0096595 .0318137 avgcumdosew1 | -.1191818 2.032868 -0.06 0.953 -4.103531 3.865167 bf4 | -.1190418 .0397853 -2.99 0.003 -.1970196 -.0410639 bf4Xd1 | .0200494 .0609254 0.33 0.742 -.0993622 .139461 ageXd1 | .0057703 .0330253 0.17 0.861 -.0589581 .0704986 radhlw1Xd1 | -.001995 .0056092 -0.36 0.722 -.0129888 .0089987 _cons | -5.587746 1.23496 -4.52 0.000 -8.008224 -3.167268 ------------------------------------------------------------------------------ Logistic model for HP2inthob -------- True -------- Classified | D ~D | Total -----------+--------------------------+----------- + | 30 11 | 41 - | 36 285 | 321 -----------+--------------------------+----------- Total | 66 296 | 362 Classified + if predicted Pr(D) >= .5 True D defined as HP2inthob != 0 -------------------------------------------------- Sensitivity Pr( +| D) 45.45% Specificity Pr( -|~D) 96.28% Positive predictive value Pr( D| +) 73.17% Negative predictive value Pr(~D| -) 88.79% -------------------------------------------------- False + rate for true ~D Pr( +|~D) 3.72% False - rate for true D Pr( -| D) 54.55% False + rate for classified + Pr(~D| +) 26.83% False - rate for classified - Pr( D| -) 11.21% -------------------------------------------------- Correctly classified 87.02% -------------------------------------------------- Logistic model for HP2inthob, goodness-of-fit test number of observations = 362 number of covariate patterns = 360 Pearson chi2(352) = 446.24 Prob > chi2 = 0.0005 Measures of Fit for logistic of HP2inthob Log-Lik Intercept Only: -171.912 Log-Lik Full Model: -126.114 D(354): 252.228 LR(7): 91.595 Prob > LR: 0.000 McFadden's R2: 0.266 McFadden's Adj R2: 0.220 Maximum Likelihood R2: 0.224 Cragg & Uhler's R2: 0.365 McKelvey and Zavoina's R2: 0.431 Efron's R2: 0.311 Variance of y*: 5.782 Variance of error: 3.290 Count R2: 0.870 Adj Count R2: 0.288 AIC: 0.741 AIC*n: 268.228 BIC: -1833.414 BIC': -50.354 1165 . scalar InthbModFw1 = "none" 1166 . 1167 . title4 " dose- interests and hobbies mediator effect models" ------------------------------------------------------------------------------- dose- interests and hobbies mediator effect models ------------------------------------------------------------------------------- 1168 . 1169 . * age is a mediating effect for males for Dose=> sex life for men 1170 . 1171 . glm age avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 1172 . glm HP2inthob age if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 224.8585 Iteration 2: deviance = 219.8355 Iteration 3: deviance = 219.6997 Iteration 4: deviance = 219.6996 Iteration 5: deviance = 219.6996 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 219.6996277 (1/df) Deviance = .6499989 Pearson = 354.7105724 (1/df) Pearson = 1.04944 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1750.484 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .031726 .0062563 5.07 0.000 .0194638 .0439882 _cons | -2.867608 .344284 -8.33 0.000 -3.542392 -2.192824 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1173 . 1174 . 1175 . des illw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- illw1 double %8.0g Total number of illnesses experienced in time period 1976-1986 1176 . glm illw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -151.48261 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = .1435742 Deviance = 48.52808533 (1/df) Deviance = .1435742 Pearson = 48.52808533 (1/df) Pearson = .1435742 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = .9028389 Log likelihood = -151.4826069 BIC = -1921.656 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0087277 .0123318 0.71 0.479 -.0154423 .0328976 _cons | .0962541 .0212201 4.54 0.000 .0546635 .1378446 ------------------------------------------------------------------------------ 1177 . glm HP2inthob illw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 237.5512 Iteration 2: deviance = 236.4934 Iteration 3: deviance = 236.4907 Iteration 4: deviance = 236.4907 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 236.4906909 (1/df) Deviance = .6996766 Pearson = 341.1245148 (1/df) Pearson = 1.009244 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1733.693 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .2627617 .1703489 1.54 0.123 -.071116 .5966393 _cons | -1.249391 .0787713 -15.86 0.000 -1.40378 -1.095002 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1178 . 1179 . * radhlw1 is a possible mediator for wave 1 males and interests and hobbies. 1180 . des radhlw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- radhlw1 double %8.0g Self-perceived Chornobyl health threat in wave 1 1181 . glm radhlw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1710.3417 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1378.645 Deviance = 465981.8893 (1/df) Deviance = 1378.645 Pearson = 465981.8893 (1/df) Pearson = 1378.645 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.0726 Log likelihood = -1710.341694 BIC = 464011.7 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 2.398285 1.208412 1.98 0.047 .0298407 4.766729 _cons | 44.66477 2.079384 21.48 0.000 40.58925 48.74029 ------------------------------------------------------------------------------ 1182 . glm HP2inthob radhlw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 212.8979 Iteration 2: deviance = 203.1654 Iteration 3: deviance = 202.4404 Iteration 4: deviance = 202.4342 Iteration 5: deviance = 202.4342 Iteration 6: deviance = 202.4342 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 202.4342472 (1/df) Deviance = .5989179 Pearson = 366.2964245 (1/df) Pearson = 1.083717 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1767.749 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0153039 .0021709 7.05 0.000 .0110489 .0195589 _cons | -2.114122 .1635196 -12.93 0.000 -2.434615 -1.79363 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1183 . 1184 . 1185 . des shfamw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 1186 . glm shfamw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1722.3986 Generalized linear models No. of obs = 339 Optimization : ML Residual df = 337 Scale parameter = 1524.9 Deviance = 513891.1861 (1/df) Deviance = 1524.9 Pearson = 513891.1861 (1/df) Pearson = 1524.9 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.17344 Log likelihood = -1722.398621 BIC = 511927.8 ------------------------------------------------------------------------------ | OIM shfamw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.8322929 1.270907 -0.65 0.513 -3.323226 1.65864 _cons | 38.72924 2.190054 17.68 0.000 34.43681 43.02167 ------------------------------------------------------------------------------ 1187 . glm HP2inthob shfamw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 236.3247 Iteration 2: deviance = 234.8343 Iteration 3: deviance = 234.8258 Iteration 4: deviance = 234.8258 Iteration 5: deviance = 234.8258 Generalized linear models No. of obs = 339 Optimization : MQL Fisher scoring Residual df = 337 (IRLS EIM) Scale parameter = 1 Deviance = 234.8258277 (1/df) Deviance = .6968125 Pearson = 340.806701 (1/df) Pearson = 1.011296 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1728.536 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shfamw1 | .0039265 .0018944 2.07 0.038 .0002136 .0076394 _cons | -1.380389 .1125461 -12.27 0.000 -1.600976 -1.159803 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1188 . 1189 . des bf5m storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf5m float %9.0g bf5m = max(0, ecprw3 - 75) * bf4m 1190 . glm bf5m avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -2272.9908 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 37743.06 Deviance = 12757154.41 (1/df) Deviance = 37743.06 Pearson = 12757154.41 (1/df) Pearson = 37743.06 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 13.3823 Log likelihood = -2272.990823 BIC = 1.28e+07 ------------------------------------------------------------------------------ | OIM bf5m | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 14.60088 6.32277 2.31 0.021 2.208475 26.99328 _cons | 107.248 10.87995 9.86 0.000 85.92368 128.5723 ------------------------------------------------------------------------------ 1191 . glm HP2inthob bf5m if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 238.9021 Iteration 2: deviance = 238.0124 Iteration 3: deviance = 238.0098 Iteration 4: deviance = 238.0098 Iteration 5: deviance = 238.0098 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 238.0098185 (1/df) Deviance = .7041711 Pearson = 339.8073754 (1/df) Pearson = 1.005347 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1732.174 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf5m | .0001615 .000378 0.43 0.669 -.0005794 .0009024 _cons | -1.236147 .0878768 -14.07 0.000 -1.408382 -1.063912 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1192 . 1193 . glm bf4 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1026.9659 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 24.75428 Deviance = 8366.946191 (1/df) Deviance = 24.75428 Pearson = 8366.946191 (1/df) Pearson = 24.75428 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.05274 Log likelihood = -1026.965868 BIC = 6396.763 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.1031788 .161925 -0.64 0.524 -.4205461 .2141884 _cons | 12.54134 .2786337 45.01 0.000 11.99523 13.08746 ------------------------------------------------------------------------------ 1194 . glm HP2inthob bf4 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 201.5914 Iteration 2: deviance = 191.8769 Iteration 3: deviance = 191.2682 Iteration 4: deviance = 191.2622 Iteration 5: deviance = 191.2622 Iteration 6: deviance = 191.2622 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 191.2622075 (1/df) Deviance = .5658645 Pearson = 305.593706 (1/df) Pearson = .9041234 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1778.921 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.1182178 .013496 -8.76 0.000 -.1446695 -.091766 _cons | .0441185 .1517879 0.29 0.771 -.2533804 .3416174 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1195 . 1196 . 1197 . scalar inthobMw1 = "age " 1198 . 1199 . * age is a mediating effect for females for Dose=> sex life for women 1200 . glm age avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 1201 . glm HP2inthob age if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 293.9671 Iteration 2: deviance = 289.27 Iteration 3: deviance = 289.1951 Iteration 4: deviance = 289.1951 Iteration 5: deviance = 289.1951 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 289.1950995 (1/df) Deviance = .8010945 Pearson = 415.3464621 (1/df) Pearson = 1.150544 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1838.684 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0515323 .0068567 7.52 0.000 .0380933 .0649713 _cons | -3.649661 .3847198 -9.49 0.000 -4.403698 -2.895624 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1202 . 1203 . * illw1 is a mediating effect for females for Dose=> sex life for women 1204 . glm illw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -259.70777 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = .2462383 Deviance = 88.89203958 (1/df) Deviance = .2462383 Pearson = 88.89203958 (1/df) Pearson = .2462383 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 1.441916 Log likelihood = -259.7077741 BIC = -2038.987 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0275426 .1585711 _cons | .1570822 .0305304 5.15 0.000 .0972436 .2169207 ------------------------------------------------------------------------------ 1205 . glm HP2inthob illw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 344.2026 Iteration 2: deviance = 344.1217 Iteration 3: deviance = 344.1216 Iteration 4: deviance = 344.1216 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 344.121648 (1/df) Deviance = .9532456 Pearson = 362.9817892 (1/df) Pearson = 1.00549 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1783.758 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .04902 .1470127 0.33 0.739 -.2391195 .3371596 _cons | -.9175059 .0797583 -11.50 0.000 -1.073829 -.7611824 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1206 . 1207 . des bf4 // soma recentered storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4 float %9.0g bf4 = max(0, 24 - BSIsoma) 1208 . * bf4 is a mediting effect for females for Dose=> sex life for women 1209 . glm bf4 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1109.0162 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 26.52082 Deviance = 9574.015672 (1/df) Deviance = 26.52082 Pearson = 9574.015672 (1/df) Pearson = 26.52082 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.121302 Log likelihood = -1109.016226 BIC = 7446.136 ------------------------------------------------------------------------------ | OIM bf4 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.508835 .4927379 -3.06 0.002 -2.474583 -.5430862 _cons | 10.99384 .3168463 34.70 0.000 10.37284 11.61485 ------------------------------------------------------------------------------ 1210 . glm HP2inthob bf4 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 304.5744 Iteration 2: deviance = 301.504 Iteration 3: deviance = 301.4488 Iteration 4: deviance = 301.4487 Iteration 5: deviance = 301.4487 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 301.4486672 (1/df) Deviance = .8350379 Pearson = 341.1451194 (1/df) Pearson = .9450003 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1826.431 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4 | -.0993749 .0142594 -6.97 0.000 -.1273229 -.071427 _cons | .0117358 .1447759 0.08 0.935 -.2720197 .2954914 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1211 . 1212 . des bf4m // soma recentered storage display value variable name type format label variable label ------------------------------------------------------------------------------- bf4m float %9.0g bf4m = max(0, 32 - BSIsoma) 1213 . * bf4m is a possible mediating effect for female sex life 1214 . glm bf4m avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1140.521 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 31.548 Deviance = 11388.82943 (1/df) Deviance = 31.548 Pearson = 11388.82943 (1/df) Pearson = 31.548 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 6.294882 Log likelihood = -1140.521046 BIC = 9260.95 ------------------------------------------------------------------------------ | OIM bf4m | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -1.563541 .5374133 -2.91 0.004 -2.616852 -.5102303 _cons | 18.82212 .3455741 54.47 0.000 18.1448 19.49943 ------------------------------------------------------------------------------ 1215 . glm HP2sxlife bf4m if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 356.1451 Iteration 2: deviance = 355.6209 Iteration 3: deviance = 355.6178 Iteration 4: deviance = 355.6178 Iteration 5: deviance = 355.6178 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 355.6178202 (1/df) Deviance = .9850909 Pearson = 340.0032816 (1/df) Pearson = .9418373 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1772.262 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- bf4m | -.0990193 .0135516 -7.31 0.000 -.12558 -.0724586 _cons | 1.063679 .244399 4.35 0.000 .584666 1.542693 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1216 . 1217 . des shfamw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shfamw1 double %8.0g Percentage of strains and hassles related to family in 1986 1218 . glm shfamw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1722.3986 Generalized linear models No. of obs = 339 Optimization : ML Residual df = 337 Scale parameter = 1524.9 Deviance = 513891.1861 (1/df) Deviance = 1524.9 Pearson = 513891.1861 (1/df) Pearson = 1524.9 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.17344 Log likelihood = -1722.398621 BIC = 511927.8 ------------------------------------------------------------------------------ | OIM shfamw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.8322929 1.270907 -0.65 0.513 -3.323226 1.65864 _cons | 38.72924 2.190054 17.68 0.000 34.43681 43.02167 ------------------------------------------------------------------------------ 1219 . glm HP2sxlife shfamw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 339.011 Iteration 2: deviance = 338.8015 Iteration 3: deviance = 338.8014 Iteration 4: deviance = 338.8014 Generalized linear models No. of obs = 339 Optimization : MQL Fisher scoring Residual df = 337 (IRLS EIM) Scale parameter = 1 Deviance = 338.8013645 (1/df) Deviance = 1.005345 Pearson = 339.2776273 (1/df) Pearson = 1.006759 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1624.561 ------------------------------------------------------------------------------ | EIM HP2sxlife | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shfamw1 | .0019797 .0019787 1.00 0.317 -.0018986 .0058579 _cons | -.9179132 .1113818 -8.24 0.000 -1.136218 -.6996088 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1220 . 1221 . des shrelaw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- shrelaw1 double %8.0g Percentage of strains and hassles related to relationships in 1986 1222 . glm shrelaw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1715.4166 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1420.421 Deviance = 480102.3567 (1/df) Deviance = 1420.421 Pearson = 480102.3567 (1/df) Pearson = 1420.421 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.10245 Log likelihood = -1715.416629 BIC = 478132.2 ------------------------------------------------------------------------------ | OIM shrelaw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | -.0649405 1.226584 -0.05 0.958 -2.469002 2.339121 _cons | 29.57199 2.110654 14.01 0.000 25.43518 33.7088 ------------------------------------------------------------------------------ 1223 . glm HP2inthob shrelaw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 238.2169 Iteration 2: deviance = 237.1515 Iteration 3: deviance = 237.1479 Iteration 4: deviance = 237.1479 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 237.1479061 (1/df) Deviance = .701621 Pearson = 340.4217715 (1/df) Pearson = 1.007165 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1733.036 ------------------------------------------------------------------------------ | EIM HP2inthob | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- shrelaw1 | .0022898 .0019458 1.18 0.239 -.0015239 .0061035 _cons | -1.289369 .0985609 -13.08 0.000 -1.482545 -1.096193 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1224 . 1225 . title4 "6. Summary matrix for dose - interests and hobbies impact" ------------------------------------------------------------------------------- 6. Summary matrix for dose - interests and hobbies impact ------------------------------------------------------------------------------- 1226 . *xx summary of mediating effects: males only age and illw1 2 1227 . *xx females: age 1228 . scalar SigDoseInthbMw1 = "no" 1229 . scalar MainEffInthbMw1 = "age radhlw1 shfamw1" 1230 . scalar InthbModMw1 = "none" 1231 . scalar InthbModFw1 = "none" 1232 . scalar SigdoseInthbFw1 = "no" 1233 . scalar MainEffInthbFw1 = "age radhlw1 bf4" 1234 . * summary of inthob moderator effects none 1235 . scalar inthobMedMw1 = "age " 1236 . scalar inthobMedFw1 = "age bf4 illw1 bf4m" 1237 . scalar SigdoseMEinthob = "no" 1238 . scalar MainEffMw1 = "radhlw1 bf4 bf40" 1239 . * no sign main dose effect for males 1240 . * no male moderators 1241 . * 3 signif main effects in male main effect model 1242 . 1243 . 1244 . * no signif dose main effect for females 1245 . * 3 main female effects 1246 . * no significant female moderators 1247 . di _skip(4) 1248 . matrix define inthbMw1 = J(1,8, 0) 1249 . matrix define inthbFw1 = J(1,8, 0) 1250 . matrix colnames inthbMw1= hypnum ptnum wave gender medsig numMAsig numModsi > g numMed 1251 . matrix colnames inthbFw1= hypnum ptnum wave gender medsig numMAsig numModsi > g numMed 1252 . matrix define inthbMw1= (1, 2, 1, 1, 0, 3, 0, 1 ) 1253 . matrix define inthbFw1= (1, 2, 1, 2, 0, 3, 0, 4 ) 1254 . matrix rowname inthbMw1 = inthbM 1255 . matrix rowname inthbFw1 = inthobF 1256 . matlist inthbMw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- inthbM | 1 2 1 1 0 3 > 0 1 1257 . matlist inthbFw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- inthobF | 1 2 1 2 0 3 > 0 4 1258 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1 \ hmcrMw1 \ hmcrFw1 \ sp > Mw1 /// > \ spFw1 \ prbfamMw1 \ prbfamFw1 \ sxlifeMw1 \ sxlifeFw1 \ inthb > Mw1 \ inthbFw1) 1259 . 1260 . matlist H1pt2w1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 1 > 1 2 r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 2 > 0 2 spMw1 | 1 2 1 1 1 2 > 2 1 spFw1 | 1 2 1 2 0 5 > 0 2 prbfamMw1 | 1 2 1 1 0 2 > 0 0 prbfamFw1 | 1 2 1 2 0 3 > 0 2 sxlifeMw1 | 1 2 1 1 0 3 > 0 1 sxlifeFw1 | 1 2 1 2 0 4 > 0 5 inthbM | 1 2 1 1 0 3 > 0 1 inthobF | 1 2 1 2 0 3 > 0 4 1261 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 1262 . matrix rownames H1pt2w1 = wkMw1 wkFw1 hmcrMw1 hmcrFw1 socprbMw1 s > ocprbFw1 prbfamMw1 prbFamFw1 sxlifeMw1 sxlifeFw1 inthbMw1 inthbFw1 1263 . matlist H1pt2w1 | hypnum ptnum wave gender medsig numMAsig > numModsig numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 hmcrMw1 | 1 2 1 1 0 2 > 0 1 hmcrFw1 | 1 2 1 2 0 2 > 0 2 socprbMw1 | 1 2 1 1 1 2 > 2 1 socprbFw1 | 1 2 1 2 0 5 > 0 2 prbfamMw1 | 1 2 1 1 0 2 > 0 0 prbFamFw1 | 1 2 1 2 0 3 > 0 2 sxlifeMw1 | 1 2 1 1 0 3 > 0 1 sxlifeFw1 | 1 2 1 2 0 4 > 0 5 inthbMw1 | 1 2 1 1 0 3 > 0 1 inthbFw1 | 1 2 1 2 0 3 > 0 4 1264 . 1265 . title "7. h1 pt2 wave 1 Dose=> vacation plans impact analysis "xxxxxxxxxxxx > xxxxxxx ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** 7. h1 pt2 wave 1 Dose=> vacation plans impact analysis ***** ***** xxxxxxxxxxxxxxxxxxx ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:18:19 ***** ******************************************************************************* ******************************************************************************* 1266 . 1267 . 1268 . 1269 . cap gen hp2vactn = HP2vacatn 1270 . 1271 . forvalues j=1/1 { 2. title " H1 pt 2 wave 1 Dose = > hp2vactn main effects models" 3. set more off 4. local w1bf bf1 bf4 bf9 bf10 bf11 bf4m bf15m bf20 bf22 bf30 bf40 5. local w1bf bf1 bf4 bf9 bf10 bf11 bf14 bf15m bf20 bf22 bf30 bf40 6. local w1bf bf1 bf4 bf9 bf10 bf11 bf14 bf15m bf20 bf22 bf30 bf40 7. di _skip(3) 8. 1272 . di as input "Male model wave 1 dose-hp2vactn moderator model " 9. di _skip(4) 10. xi: logistic hp2vactn age i.educ occ1w`j'-occ8w`j' /// > marrw`j'1- marrw`j'3 marrw`j'5-marrw`j'6 inc1w`j'-inc4w`j' /// > radhlw`j' havmil avgcumdosew`j' `w`j'bf' /// > deaw`j' dvcew`j' sepaw`j' accdw`j' movew`j' /// > illw`j' shfamw`j' shhlw`j' shjobw`j' shrelaw`j' suprtw`j' suchrw`j' havmils > q /// > radhlw`j' avgcumdosew`j' if gender==1, coef 11. di _skip(4) 12. title3 "trimmed hp2vactn main effects models for H1 no direct dose effect > for male" 13. pwcorr hp2hmcare age deaw1 shjobw1 bf7m shjobw1 havmilsq /// > radhlw1 avgcumdosew1 if gender==1, sig obs sidak star(.05) listwise 14. di _skip(1) 15. di as input "For males hp2vactn wave3 and d1 is not signif " 16. di _skip(1) 17. logistic hp2vactn age deaw1 shjobw1 bf7m havmilsq /// > radhlw1 avgcumdosew1 if /// > gender==1, coef 18. } ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** H1 pt 2 wave 1 Dose = > hp2vactn main effects models ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:18:19 ***** ******************************************************************************* ******************************************************************************* Male model wave 1 dose-hp2vactn moderator model i.educ _Ieduc_1-8 (naturally coded; _Ieduc_1 omitted) note: _Ieduc_4 != 0 predicts failure perfectly _Ieduc_4 dropped and 12 obs not used note: _Ieduc_7 != 0 predicts failure perfectly _Ieduc_7 dropped and 4 obs not used note: _Ieduc_8 != 0 predicts failure perfectly _Ieduc_8 dropped and 2 obs not used note: occ6w1 != 0 predicts failure perfectly occ6w1 dropped and 4 obs not used note: occ7w1 != 0 predicts failure perfectly occ7w1 dropped and 4 obs not used note: marrw15 != 0 predicts success perfectly marrw15 dropped and 1 obs not used note: bf15m != 0 predicts failure perfectly bf15m dropped and 19 obs not used note: dvcew1 != 0 predicts failure perfectly dvcew1 dropped and 2 obs not used note: sepaw1 != 0 predicts success perfectly sepaw1 dropped and 1 obs not used note: _Ieduc_6 omitted because of collinearity note: marrw16 omitted because of collinearity note: radhlw1 omitted because of collinearity note: avgcumdosew1 omitted because of collinearity Logistic regression Number of obs = 284 LR chi2(41) = 101.69 Prob > chi2 = 0.0000 Log likelihood = -57.124451 Pseudo R2 = 0.4709 ------------------------------------------------------------------------------ hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .1177419 .0451828 2.61 0.009 .0291854 .2062985 _Ieduc_2 | -3.055944 1.716937 -1.78 0.075 -6.421078 .3091893 _Ieduc_3 | -.212761 .8269168 -0.26 0.797 -1.833488 1.407966 _Ieduc_4 | 0 (omitted) _Ieduc_5 | -.158373 .8405787 -0.19 0.851 -1.805877 1.489131 _Ieduc_6 | 0 (omitted) _Ieduc_7 | 0 (omitted) _Ieduc_8 | 0 (omitted) occ1w1 | .9327826 3.048048 0.31 0.760 -5.041281 6.906846 occ2w1 | 1.600773 3.127281 0.51 0.609 -4.528585 7.730132 occ3w1 | .8629517 3.222587 0.27 0.789 -5.453202 7.179105 occ4w1 | .7047753 3.193274 0.22 0.825 -5.553927 6.963477 occ5w1 | .9185675 3.369728 0.27 0.785 -5.685978 7.523113 occ6w1 | 0 (omitted) occ7w1 | 0 (omitted) occ8w1 | 2.576584 3.31921 0.78 0.438 -3.928947 9.082116 marrw11 | 8.063251 1454.164 0.01 0.996 -2842.047 2858.173 marrw12 | 8.396586 1454.165 0.01 0.995 -2841.715 2858.509 marrw13 | 6.592655 1454.165 0.00 0.996 -2843.518 2856.703 marrw15 | 0 (omitted) marrw16 | 0 (omitted) inc1w1 | 1.72241 3.405241 0.51 0.613 -4.951739 8.396559 inc2w1 | 1.681005 3.353329 0.50 0.616 -4.891398 8.253409 inc3w1 | 1.742795 3.224514 0.54 0.589 -4.577137 8.062726 inc4w1 | .22304 3.501992 0.06 0.949 -6.640738 7.086818 radhlw1 | .0177143 .0102482 1.73 0.084 -.0023719 .0378004 havmil | -.000285 .0120623 -0.02 0.981 -.0239267 .0233567 avgcumdosew1 | -.101335 .173417 -0.58 0.559 -.441226 .238556 bf1 | -.067538 .109717 -0.62 0.538 -.2825793 .1475033 bf4 | -.2732381 .0860278 -3.18 0.001 -.4418495 -.1046267 bf9 | .045179 .0491094 0.92 0.358 -.0510737 .1414317 bf10 | -.0859725 .0443879 -1.94 0.053 -.1729712 .0010262 bf11 | .0439137 .1246785 0.35 0.725 -.2004517 .288279 bf14 | -.0000515 .0001058 -0.49 0.627 -.0002589 .0001559 bf15m | 0 (omitted) bf20 | .0464251 .1033689 0.45 0.653 -.1561741 .2490244 bf22 | .0001389 .0001358 1.02 0.306 -.0001273 .0004051 bf30 | -.001371 .0006346 -2.16 0.031 -.0026148 -.0001273 bf40 | .3396411 .2165949 1.57 0.117 -.0848772 .7641593 deaw1 | -.5911018 .7169403 -0.82 0.410 -1.996279 .8140753 dvcew1 | 0 (omitted) sepaw1 | 0 (omitted) accdw1 | 1.690611 1.483962 1.14 0.255 -1.2179 4.599123 movew1 | .7972512 .7853957 1.02 0.310 -.7420961 2.336598 illw1 | -1.434715 .7967422 -1.80 0.072 -2.996301 .1268714 shfamw1 | -.0116324 .0105102 -1.11 0.268 -.032232 .0089671 shhlw1 | .0339732 .0188557 1.80 0.072 -.0029832 .0709297 shjobw1 | .0121109 .0125268 0.97 0.334 -.0124412 .036663 shrelaw1 | -.0146336 .0112271 -1.30 0.192 -.0366384 .0073711 suprtw1 | .1086712 .0510741 2.13 0.033 .0085679 .2087746 suchrw1 | -.0573692 .0444525 -1.29 0.197 -.1444944 .0297561 havmilsq | -6.00e-06 .0000223 -0.27 0.787 -.0000496 .0000376 radhlw1 | 0 (omitted) avgcumdosew1 | 0 (omitted) _cons | -21.03776 1454.176 -0.01 0.988 -2871.17 2829.095 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. ------------------------------------------------------------------------------- title3 : trimmed hp2vactn main effects models for H1 no direct dose effect for > male 18 Jun 2012 18:18:21 computer Macintosh (Intel 64-bit) folder /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/ > h1pt2 Data file chwide16june2012.dta currrently has 2388 variables and 703 obse > rvations | hp2hmc~e age deaw1 shjobw1 bf7m shjobw1 havmilsq -------------+--------------------------------------------------------------- hp2hmcare | 1.0000 | | 340 | age | 0.2761* 1.0000 | 0.0000 | 340 340 | deaw1 | 0.0095 0.2204* 1.0000 | 1.0000 0.0015 | 340 340 340 | shjobw1 | 0.2490* 0.2162* -0.0439 1.0000 | 0.0001 0.0021 1.0000 | 340 340 340 340 | bf7m | -0.1339 -0.0182 0.1915* -0.2097* 1.0000 | 0.3860 1.0000 0.0137 0.0035 | 340 340 340 340 340 | shjobw1 | 0.2490* 0.2162* -0.0439 1.0000* -0.2097* 1.0000 | 0.0001 0.0021 1.0000 0.0000 0.0035 | 340 340 340 340 340 340 | havmilsq | -0.0347 0.0207 -0.0393 0.0027 -0.0420 0.0027 1.0000 | 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 | 340 340 340 340 340 340 340 | radhlw1 | 0.2183* 0.3167* 0.0077 0.4115* 0.0528 0.4115* -0.1218 | 0.0018 0.0000 1.0000 0.0000 1.0000 0.0000 0.5939 | 340 340 340 340 340 340 340 | avgcumdosew1 | 0.0010 0.0918 -0.0226 0.0656 0.0206 0.0656 -0.0345 | 1.0000 0.9680 1.0000 0.9999 1.0000 0.9999 1.0000 | 340 340 340 340 340 340 340 | | radhlw1 avgcum~1 -------------+------------------ radhlw1 | 1.0000 | | 340 | avgcumdosew1 | 0.1073 1.0000 | 0.8298 | 340 340 | For males hp2vactn wave3 and d1 is not signif Logistic regression Number of obs = 340 LR chi2(7) = 58.56 Prob > chi2 = 0.0000 Log likelihood = -95.872164 Pseudo R2 = 0.2340 ------------------------------------------------------------------------------ hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0698089 .0169544 4.12 0.000 .0365789 .1030389 deaw1 | -1.031518 .5915573 -1.74 0.081 -2.190949 .1279133 shjobw1 | .0185312 .0064023 2.89 0.004 .005983 .0310794 bf7m | .0000825 .0002766 0.30 0.766 -.0004597 .0006247 havmilsq | -6.78e-06 8.13e-06 -0.83 0.405 -.0000227 9.17e-06 radhlw1 | .0131121 .005597 2.34 0.019 .0021422 .0240821 avgcumdosew1 | -.1161641 .1842657 -0.63 0.528 -.4773183 .2449901 _cons | -7.463166 1.17919 -6.33 0.000 -9.774335 -5.151996 ------------------------------------------------------------------------------ 1273 . 1274 . scalar SigDoseVactnMw1 = "no" 1275 . scalar MainEffVactnMw1 = "age radhlw1 shjobw1 " 1276 . 1277 . local cn7:colnames(e(b)) 1278 . di "`cn7'" age deaw1 shjobw1 bf7m havmilsq radhlw1 avgcumdosew1 _cons 1279 . local len7 = length("`cn7'") 1280 . di `len7' 58 1281 . local len7b = `len7' - 6 1282 . di `len7b' 52 1283 . local myvarlist = substr("`cn7'",1,`len7b') 1284 . di "`myvarlist'" age deaw1 shjobw1 bf7m havmilsq radhlw1 avgcumdosew1 1285 . 1286 . foreach var in `myvarlist' { 2. cap gen `var'Xd1 = `var'*avgcumdosew1 3. } 1287 . 1288 . title " Trimmed male main effects dose=> vacation plans model" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Trimmed male main effects dose=> vacation plans model ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:18:22 ***** ******************************************************************************* ******************************************************************************* 1289 . di as input "No sig main male dose main effects model" No sig main male dose main effects model 1290 . sw, pr(.1): logit hp2vacatn `myvarlist' if gender==1 begin with full model p = 0.7655 >= 0.1000 removing bf7m p = 0.5322 >= 0.1000 removing avgcumdosew1 p = 0.4214 >= 0.1000 removing havmilsq Logistic regression Number of obs = 340 LR chi2(4) = 56.62 Prob > chi2 = 0.0000 Log likelihood = -96.8403 Pseudo R2 = 0.2262 ------------------------------------------------------------------------------ hp2vacatn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0683746 .0168652 4.05 0.000 .0353194 .1014298 deaw1 | -.9608875 .5756492 -1.67 0.095 -2.089139 .1673642 shjobw1 | .0175447 .0060455 2.90 0.004 .0056958 .0293937 radhlw1 | .013369 .0055515 2.41 0.016 .0024883 .0242496 _cons | -7.378809 1.054557 -7.00 0.000 -9.445702 -5.311916 ------------------------------------------------------------------------------ 1291 . 1292 . 1293 . local cn8:colnames(e(b)) 1294 . di "`cn8'" age deaw1 shjobw1 radhlw1 _cons 1295 . local len8 = length("`cn8'") 1296 . di `len7' 58 1297 . local len8b = `len8' - 6 1298 . di `len8b' 25 1299 . local myvarlist = substr("`cn8'",1,`len8b') 1300 . di "`myvarlist'" age deaw1 shjobw1 radhlw1 1301 . 1302 . logit hp2vacatn age radhlw1 avgcumdosew1 ageXd1 radhlw1Xd1 if gender==1 Iteration 0: log likelihood = -125.15243 Iteration 1: log likelihood = -107.61459 Iteration 2: log likelihood = -104.56768 Iteration 3: log likelihood = -104.54823 Iteration 4: log likelihood = -104.54822 Logistic regression Number of obs = 340 LR chi2(5) = 41.21 Prob > chi2 = 0.0000 Log likelihood = -104.54822 Pseudo R2 = 0.1646 ------------------------------------------------------------------------------ hp2vacatn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0601195 .0209224 2.87 0.004 .0191123 .1011267 radhlw1 | .0187374 .0058584 3.20 0.001 .0072551 .0302197 avgcumdosew1 | .1429635 3.450929 0.04 0.967 -6.620734 6.906661 ageXd1 | -.0052571 .0512646 -0.10 0.918 -.1057338 .0952196 radhlw1Xd1 | .0008937 .0086815 0.10 0.918 -.0161217 .0179091 _cons | -6.233448 1.294444 -4.82 0.000 -8.770513 -3.696384 ------------------------------------------------------------------------------ 1303 . 1304 . title "Trimmed male wave 1 interaction dose=> vacation plans model" ******************************************************************************* ******************************************************************************* ***** ***** ***** ***** ***** Trimmed male wave 1 interaction dose=> vacation plans model ***** ***** ***** ***** ***** ***** 18 Jun 2012 18:18:30 ***** ******************************************************************************* ******************************************************************************* 1305 . logit hp2vacatn age radhlw1 ageXd1 bf7m avgcumdosew1 bf4m bf4mXd1 bf7mXd1 if > gender==1 Iteration 0: log likelihood = -125.15243 Iteration 1: log likelihood = -102.74095 Iteration 2: log likelihood = -94.448542 Iteration 3: log likelihood = -93.949822 Iteration 4: log likelihood = -93.920677 Iteration 5: log likelihood = -93.920292 Iteration 6: log likelihood = -93.92029 Logistic regression Number of obs = 340 LR chi2(8) = 62.46 Prob > chi2 = 0.0000 Log likelihood = -93.92029 Pseudo R2 = 0.2496 ------------------------------------------------------------------------------ hp2vacatn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0643459 .0308181 2.09 0.037 .0039435 .1247483 radhlw1 | .0089983 .0057782 1.56 0.119 -.0023267 .0203233 ageXd1 | -.0876104 .0995184 -0.88 0.379 -.2826628 .107442 bf7m | .0001157 .0003967 0.29 0.771 -.0006618 .0008932 avgcumdosew1 | 7.211044 8.155585 0.88 0.377 -8.773609 23.1957 bf4m | -.1044109 .0615606 -1.70 0.090 -.2250674 .0162456 bf4mXd1 | -.1652155 .1615499 -1.02 0.306 -.4818476 .1514165 bf7mXd1 | .0005521 .0007567 0.73 0.466 -.000931 .0020353 _cons | -4.176182 2.553316 -1.64 0.102 -9.18059 .8282258 ------------------------------------------------------------------------------ Note: 1 failure and 0 successes completely determined. 1306 . 1307 . scalar vactnModMw1 ="none" 1308 . 1309 . title4 "Trimmed Female model wave 1 main effects dose-hp2vacatn model " ------------------------------------------------------------------------------- Trimmed Female model wave 1 main effects dose-hp2vacatn model ------------------------------------------------------------------------------- 1310 . forvalues j=1/1 { 2. local w1bf bf1 bf4 bf9 bf10 bf11 bf14m bf15m bf20 bf22 bf30 bf40 3. 1311 . xi: logistic hp2vacatn age radhlw`j' avgcumdosew`j' /// > deaw`j' suchrw`j' /// > if gender==2, coef difficult iterate(50) 4. 1312 . } Logistic regression Number of obs = 359 LR chi2(5) = 79.84 Prob > chi2 = 0.0000 Log likelihood = -126.83022 Pseudo R2 = 0.2394 ------------------------------------------------------------------------------ hp2vacatn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0851484 .0156032 5.46 0.000 .0545668 .11573 radhlw1 | .0210852 .0051434 4.10 0.000 .0110043 .031166 avgcumdosew1 | .097081 .2416199 0.40 0.688 -.3764853 .5706473 deaw1 | .0184403 .1587264 0.12 0.908 -.2926576 .3295383 suchrw1 | .0205099 .0120186 1.71 0.088 -.0030461 .0440658 _cons | -7.695923 .966869 -7.96 0.000 -9.590951 -5.800894 ------------------------------------------------------------------------------ 1313 . 1314 . sw, pr(.1): logit hp2vactn age radhlw1 avgcumdosew1 havmilsq ageXd1 radhlw1Xd > 1 if gender==2 begin with full model p = 0.8147 >= 0.1000 removing havmilsq p = 0.7305 >= 0.1000 removing avgcumdosew1 p = 0.4968 >= 0.1000 removing ageXd1 p = 0.4278 >= 0.1000 removing radhlw1Xd1 Logistic regression Number of obs = 362 LR chi2(2) = 75.57 Prob > chi2 = 0.0000 Log likelihood = -129.53941 Pseudo R2 = 0.2258 ------------------------------------------------------------------------------ hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0865333 .0154134 5.61 0.000 .0563237 .116743 radhlw1 | .0216968 .0050096 4.33 0.000 .0118782 .0315154 _cons | -7.711441 .9591085 -8.04 0.000 -9.591259 -5.831623 ------------------------------------------------------------------------------ 1315 . 1316 . 1317 . 1318 . scalar SigDoseVactnMw1 = "no" 1319 . scalar MainEffVactnMw1 = "age radhlw1" 1320 . scalar VactnModMw1 = "none" 1321 . 1322 . * summary of male moderating effects: no sign main dose effect in main effec > ts model 1323 . * no signif male moderators 1324 . * 3 significant main effects in main effects model 1325 . 1326 . * summary of female moderation main effects: no signif main dose effect 1327 . 1328 . 1329 . scalar SigDoseVactnFw1 = "no" 1330 . scalar MainEffVactnFw1 = "age radhlw1 bf7m" 1331 . 1332 . cap gen suchrw1Xd1 = suchrw1*avgcumdosew1 1333 . 1334 . 1335 . scalar VacatnModFw1 = "none" 1336 . *xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx > xxx 1337 . cap gen hp2vactn = HP2vacatn 1338 . title4 "Male mediator tests for vacation plans impact of dose" ------------------------------------------------------------------------------- Male mediator tests for vacation plans impact of dose ------------------------------------------------------------------------------- 1339 . * for males 1340 . 1341 . * age is a mediator for males 1342 . glm age avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1331.608 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 148.5632 Deviance = 50214.37624 (1/df) Deviance = 148.5632 Pearson = 50214.37624 (1/df) Pearson = 148.5632 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.844753 Log likelihood = -1331.607976 BIC = 48244.19 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .6719789 .3966839 1.69 0.090 -.1055072 1.449465 _cons | 48.89394 .6825967 71.63 0.000 47.55607 50.2318 ------------------------------------------------------------------------------ 1343 . glm hp2vactn age if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 230.5483 Iteration 2: deviance = 224.6254 Iteration 3: deviance = 224.4149 Iteration 4: deviance = 224.4147 Iteration 5: deviance = 224.4147 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 224.4146771 (1/df) Deviance = .6639487 Pearson = 345.2370578 (1/df) Pearson = 1.021411 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1745.769 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0377399 .0063833 5.91 0.000 .0252289 .0502509 _cons | -3.150096 .3549593 -8.87 0.000 -3.845803 -2.454388 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1344 . 1345 . 1346 . des illw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- illw1 double %8.0g Total number of illnesses experienced in time period 1976-1986 1347 . glm illw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -151.48261 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = .1435742 Deviance = 48.52808533 (1/df) Deviance = .1435742 Pearson = 48.52808533 (1/df) Pearson = .1435742 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = .9028389 Log likelihood = -151.4826069 BIC = -1921.656 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0087277 .0123318 0.71 0.479 -.0154423 .0328976 _cons | .0962541 .0212201 4.54 0.000 .0546635 .1378446 ------------------------------------------------------------------------------ 1348 . glm hp2vactn illw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 250.8272 Iteration 2: deviance = 250.3035 Iteration 3: deviance = 250.3029 Iteration 4: deviance = 250.3029 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 250.3028927 (1/df) Deviance = .7405411 Pearson = 339.9984851 (1/df) Pearson = 1.005913 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1719.881 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | -.0103789 .2018135 -0.05 0.959 -.4059261 .3851682 _cons | -1.171022 .0782871 -14.96 0.000 -1.324462 -1.017583 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1349 . 1350 . des radhlw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- radhlw1 double %8.0g Self-perceived Chornobyl health threat in wave 1 1351 . glm radhlw1 avgcumdosew1 if gender==1, fam(gaus) link(identity) Iteration 0: log likelihood = -1710.3417 Generalized linear models No. of obs = 340 Optimization : ML Residual df = 338 Scale parameter = 1378.645 Deviance = 465981.8893 (1/df) Deviance = 1378.645 Pearson = 465981.8893 (1/df) Pearson = 1378.645 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.0726 Log likelihood = -1710.341694 BIC = 464011.7 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 2.398285 1.208412 1.98 0.047 .0298407 4.766729 _cons | 44.66477 2.079384 21.48 0.000 40.58925 48.74029 ------------------------------------------------------------------------------ 1352 . glm hp2vactn radhlw1 if gender==1, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 230.8439 Iteration 2: deviance = 224.8904 Iteration 3: deviance = 224.6643 Iteration 4: deviance = 224.6638 Iteration 5: deviance = 224.6638 Generalized linear models No. of obs = 340 Optimization : MQL Fisher scoring Residual df = 338 (IRLS EIM) Scale parameter = 1 Deviance = 224.6638339 (1/df) Deviance = .6646859 Pearson = 351.6837556 (1/df) Pearson = 1.040484 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1745.52 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0123039 .0020846 5.90 0.000 .0082183 .0163896 _cons | -1.857875 .1493813 -12.44 0.000 -2.150657 -1.565093 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1353 . 1354 . * for females 1355 . 1356 . * age is a mediator for females 1357 . glm age avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1406.325 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = 136.455 Deviance = 49260.25928 (1/df) Deviance = 136.455 Pearson = 49260.25928 (1/df) Pearson = 136.455 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 7.759366 Log likelihood = -1406.325011 BIC = 47132.38 ------------------------------------------------------------------------------ | OIM age | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.973879 1.117679 3.56 0.000 1.783267 6.16449 _cons | 48.88157 .7187038 68.01 0.000 47.47293 50.2902 ------------------------------------------------------------------------------ 1358 . glm hp2vactn age if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 288.2228 Iteration 2: deviance = 282.9212 Iteration 3: deviance = 282.8 Iteration 4: deviance = 282.8 Iteration 5: deviance = 282.8 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 282.7999835 (1/df) Deviance = .7833795 Pearson = 396.9154874 (1/df) Pearson = 1.099489 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1845.079 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0510238 .0068535 7.44 0.000 .0375912 .0644564 _cons | -3.660305 .3855366 -9.49 0.000 -4.415943 -2.904667 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1359 . 1360 . * illness is a mediating effect for females = > vacatn 1361 . des illw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- illw1 double %8.0g Total number of illnesses experienced in time period 1976-1986 1362 . glm illw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -259.70777 Generalized linear models No. of obs = 363 Optimization : ML Residual df = 361 Scale parameter = .2462383 Deviance = 88.89203958 (1/df) Deviance = .2462383 Pearson = 88.89203958 (1/df) Pearson = .2462383 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 1.441916 Log likelihood = -259.7077741 BIC = -2038.987 ------------------------------------------------------------------------------ | OIM illw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | .0655142 .0474789 1.38 0.168 -.0275426 .1585711 _cons | .1570822 .0305304 5.15 0.000 .0972436 .2169207 ------------------------------------------------------------------------------ 1363 . glm hp2vactn illw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 334.5587 Iteration 2: deviance = 334.5192 Iteration 3: deviance = 334.5191 Iteration 4: deviance = 334.5191 Generalized linear models No. of obs = 363 Optimization : MQL Fisher scoring Residual df = 361 (IRLS EIM) Scale parameter = 1 Deviance = 334.5191456 (1/df) Deviance = .9266458 Pearson = 363.294411 (1/df) Pearson = 1.006356 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1793.36 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- illw1 | .1050277 .1426608 0.74 0.462 -.1745823 .3846376 _cons | -.9602112 .0797688 -12.04 0.000 -1.116555 -.8038673 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1364 . 1365 . * radhlw1 is a mediating effect for females => vactn 1366 . des radhlw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- radhlw1 double %8.0g Self-perceived Chornobyl health threat in wave 1 1367 . glm radhlw1 avgcumdosew1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1821.9477 Generalized linear models No. of obs = 362 Optimization : ML Residual df = 360 Scale parameter = 1385.301 Deviance = 498708.3025 (1/df) Deviance = 1385.301 Pearson = 498708.3025 (1/df) Pearson = 1385.301 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.07706 Log likelihood = -1821.947718 BIC = 496587.3 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 3.789972 3.562254 1.06 0.287 -3.191917 10.77186 _cons | 55.44948 2.293757 24.17 0.000 50.9538 59.94516 ------------------------------------------------------------------------------ 1368 . glm hp2vactn radhlw1 if gender==2, fam(bin) irls scale(dev) link(probit) Iteration 1: deviance = 302.9154 Iteration 2: deviance = 298.4828 Iteration 3: deviance = 298.355 Iteration 4: deviance = 298.3548 Iteration 5: deviance = 298.3548 Generalized linear models No. of obs = 362 Optimization : MQL Fisher scoring Residual df = 360 (IRLS EIM) Scale parameter = 1 Deviance = 298.3548116 (1/df) Deviance = .8287634 Pearson = 355.6125805 (1/df) Pearson = .9878127 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1822.637 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .0139561 .0022604 6.17 0.000 .0095258 .0183864 _cons | -1.855913 .1776242 -10.45 0.000 -2.20405 -1.507776 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1369 . 1370 . * summary of male moderating effects: no sign main dose effect in main effec > ts model 1371 . * no signif male moderators 1372 . * 3 significant main effects in main effects model 1373 . * summary omnibus model 1374 . des radhlw1 storage display value variable name type format label variable label ------------------------------------------------------------------------------- radhlw1 double %8.0g Self-perceived Chornobyl health threat in wave 1 1375 . glm radhlw1 avgcumdosew1 illw1 if gender==2, fam(gaus) link(identity) Iteration 0: log likelihood = -1818.8802 Generalized linear models No. of obs = 362 Optimization : ML Residual df = 359 Scale parameter = 1365.815 Deviance = 490327.6005 (1/df) Deviance = 1365.815 Pearson = 490327.6005 (1/df) Pearson = 1365.815 Variance function: V(u) = 1 [Gaussian] Link function : g(u) = u [Identity] AIC = 10.06564 Log likelihood = -1818.880199 BIC = 488212.5 ------------------------------------------------------------------------------ | OIM radhlw1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- avgcumdosew1 | 4.422493 3.546316 1.25 0.212 -2.528159 11.37315 illw1 | -9.71121 3.920388 -2.48 0.013 -17.39503 -2.02739 _cons | 56.98053 2.359945 24.14 0.000 52.35513 61.60594 ------------------------------------------------------------------------------ 1376 . glm hp2vactn radhlw1 illw1 avgcumdosew1 if gender==2, fam(bin) irls scale(dev > ) link(probit) Iteration 1: deviance = 300.2452 Iteration 2: deviance = 295.4656 Iteration 3: deviance = 295.3193 Iteration 4: deviance = 295.319 Iteration 5: deviance = 295.319 Generalized linear models No. of obs = 362 Optimization : MQL Fisher scoring Residual df = 358 (IRLS EIM) Scale parameter = 1 Deviance = 295.3190068 (1/df) Deviance = .8249134 Pearson = 359.0506546 (1/df) Pearson = 1.002935 Variance function: V(u) = u*(1-u) [Bernoulli] Link function : g(u) = invnorm(u) [Probit] BIC = -1813.89 ------------------------------------------------------------------------------ | EIM hp2vactn | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- radhlw1 | .014364 .0022973 6.25 0.000 .0098614 .0188667 illw1 | .2268401 .147174 1.54 0.123 -.0616157 .5152959 avgcumdosew1 | .1252934 .1239292 1.01 0.312 -.1176035 .3681902 _cons | -1.970546 .1907882 -10.33 0.000 -2.344484 -1.596608 ------------------------------------------------------------------------------ (Standard errors scaled using square root of deviance-based dispersion.) 1377 . 1378 . scalar SigDoseVactnMw1 = "no" 1379 . scalar MainEffVactnMw1 = "age radhlw1 shjobw1 " 1380 . scalar VactnMedMw1 = "age" 1381 . scalar VactnMedFw1 = "age illw1 radhlw1" 1382 . scalar SigDoseVactnMw1 = "no" 1383 . scalar MainEffVactnMw1 = "age radhlw1" 1384 . scalar VactnModMw1 = "none" 1385 . scalar SigDoseVactnMw1 = "no" 1386 . scalar MainEffVactnMw1 = "age radhlw1" 1387 . scalar VactnModMw1 = "none" 1388 . * Part 2 Nottingham subscales 1389 . * 1: hp2work 1390 . * 2: hp2hmcare 1391 . * 3: hp2probsoc 1392 . * 4: hp2pbfhm 1393 . * 5: hp2sexlife 1394 . * 6: hp2inthob 1395 . * 7: hp2vacatn 1396 . 1397 . *xx summary of moderator effects for females: 1398 . * no signif main dose effect 1399 . * 3 signif main effects in main effect model 1400 . * 1 moderator: deaw1Xd1 1401 . title4 "7. Summary Matrix construction of dose - vacatn plans impact" ------------------------------------------------------------------------------- 7. Summary Matrix construction of dose - vacatn plans impact ------------------------------------------------------------------------------- 1402 . 1403 . matrix define vactnMw1 = J(1,8, 0) 1404 . matrix define vactnFw1 = J(1,8, 0) 1405 . matrix colnames vactnMw1= hypnum ptnum wave gender medsig numMAsig numModsi > g numMed 1406 . matrix colnames vactnFw1= hypnum ptnum wave gender medsig numMAsig numModsi > g numMed 1407 . matrix define vactnMw1= (1, 2, 1, 1, 0, 2, 0, 1 ) 1408 . matrix define vactnFw1= (1, 2, 1, 2, 0, 3, 0, 3 ) 1409 . matrix rowname vactnMw1 = vactnM 1410 . matrix rowname vactnFw1 = vactnF 1411 . matlist vactnMw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- vactnM | 1 2 1 1 0 2 > 0 1 1412 . matlist vactnFw1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- vactnF | 1 2 1 2 0 3 > 0 3 1413 . matrix define H1pt2w1 = ( wkMw1 \ wkFw1 \ hmcrMw1 \ hmcrFw1 \ sp > Mw1 /// > \ spFw1 \ prbfamMw1 \ prbfamFw1 \ sxlifeMw1 \ sxlifeFw1 \ inthbM > w1 \ inthbFw1 \ vactnMw1 \ vactnFw1 ) 1414 . 1415 . matlist H1pt2w1 | c1 c2 c3 c4 c5 c6 > c7 c8 -------------+----------------------------------------------------------------- ----------------------- r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 1 > 1 2 r1 | 1 2 1 1 0 2 > 0 1 r1 | 1 2 1 2 0 2 > 0 2 spMw1 | 1 2 1 1 1 2 > 2 1 spFw1 | 1 2 1 2 0 5 > 0 2 prbfamMw1 | 1 2 1 1 0 2 > 0 0 prbfamFw1 | 1 2 1 2 0 3 > 0 2 sxlifeMw1 | 1 2 1 1 0 3 > 0 1 sxlifeFw1 | 1 2 1 2 0 4 > 0 5 inthbM | 1 2 1 1 0 3 > 0 1 inthobF | 1 2 1 2 0 3 > 0 4 vactnM | 1 2 1 1 0 2 > 0 1 vactnF | 1 2 1 2 0 3 > 0 3 1416 . matrix colnames H1pt2w1 = hypnum ptnum wave gender medsig numMAsig numMo > dsig numMed 1417 . matrix rownames H1pt2w1 = wkMw1 wkFw1 hmcrMw1 hmcrFw1 socprbMw1 s > ocprbFw1 prbfamhmMw1 prbfamhmFw1 sxlifeMw1 sxlifeFw1 inthbMw1 inthbFw1 vact > nMw1 vacatnFw1 1418 . matlist H1pt2w1 | hypnum ptnum wave gender medsig numMAsig > numModsig numMed -------------+----------------------------------------------------------------- ----------------------- wkMw1 | 1 2 1 1 0 2 > 0 1 wkFw1 | 1 2 1 2 0 1 > 1 2 hmcrMw1 | 1 2 1 1 0 2 > 0 1 hmcrFw1 | 1 2 1 2 0 2 > 0 2 socprbMw1 | 1 2 1 1 1 2 > 2 1 socprbFw1 | 1 2 1 2 0 5 > 0 2 prbfamhmMw1 | 1 2 1 1 0 2 > 0 0 prbfamhmFw1 | 1 2 1 2 0 3 > 0 2 sxlifeMw1 | 1 2 1 1 0 3 > 0 1 sxlifeFw1 | 1 2 1 2 0 4 > 0 5 inthbMw1 | 1 2 1 1 0 3 > 0 1 inthbFw1 | 1 2 1 2 0 3 > 0 4 vactnMw1 | 1 2 1 1 0 2 > 0 1 vacatnFw1 | 1 2 1 2 0 3 > 0 3 1419 . set more off 1420 . scalar list MainEffVactnMw1 = age radhlw1 sxlifeMedMw1 = radhlw1 MainEffsxlifeFw1 = age bf4 bf4m SigDoseSxlifeFw1 = no MainEffhmcrFw1 = age MainEffhmcrMw1 = age hmcrMedFw1 = age bf4 hmcrMedMw1 = radhlw1 MainEffwkFw1 = age MainEffwkMw1 = age VactnMedFw1 = age illw1 radhlw1 VactnMedMw1 = age VacatnModFw1 = none MainEffVactnFw1 = age radhlw1 bf7m SigDoseVactnFw1 = no VactnModMw1 = none inthobMedFw1 = age bf4 illw1 bf4m inthobMedMw1 = age inthobMw1 = age InthbModFw1 = none MainEffInthbFw1 = age radhlw1 bf4 SigdoseInthbFw1 = no InthbModMw1 = none MainEffInthbMw1 = age radhlw1 shfamw1 SigDoseInthbMw1 = no MainEffMw1 = radhlw1 bf4 bf40 sxlifeMedFw1 = age illw1 radhlw1 bf4 bf4m SigDosesxlifeMw1 = no MainEffsxlifeMw1 = age bf4 bf40 SigDosePrbfmhmMw1 = no vactnModMw1 = none SigDoseVactnMw1 = no SxLifeModFw1 = no sxlifeModFw1 = none sxlifeModMw1 = none MaineffhmcrMw1 = age bf4 bf40 SigDoseMEhmcrW1 = no PrbfmhmMedFw1 = age bf4 PrbfmhmMedMw1 = age MainEffPrbfmhmFw1 = age radhlw1 bf4 MainEffPrbfmhmMw1 = age bf4 PrbfmhmModFw1 = none PrbfmhmModMw1 = none SigDosePrbfmhmFw1 = no SigDosePrbfhmMw1 = no MainEffPrbfhmMw1 = age bf4 MainEffVactnMw2 = age radhlw2 sxlifeMedMw2 = age illw2 SigDoseSxlifeFw2 = no MainEffsxlifeFw2 = age radhlw2 bf4 bf4m MainEffPrbsocMw2 = age radhlw2 shjobw2 MainEffhmcrFw2 = age hmcrMedFw2 = age bf4 MainEffwkFw2 = age MainEffwkMw2 = age MainEffPrbsocMw1 = age bf4m SigdoseMw1 = no ProbsocMedFw1 = age bf4 ProbsocMedMw1 = radhlw1 ProbsocModFw1 = none SigDoseProbsocMw1 = no hmcrmedMw1 = radhlw1 hmcrmedFw1 = age b4 b40 SigdosehmcrFw1 = no MainEffProbSocFw1 = age radhlw1 avgcumdosew1 shrelaw1 bf4 SigDoseProbsocFw1 = yes PrbsocModMw1 = shjobw1Xd1 shrelaw1Xd1 WkhmcrMw1 = age b4 WkModFw1 = ageXd1 hmcareMedFw1 = age illw1 hmcareMedMw1 = age SigDosehmcrFw1 = no wkMedMw1 = bf40 hmcrModFw1 = none SigDoseHmcrFw1 = yes WkMedMw1 = bf40 hmcrModMw1 = none SigDosehmcrMw1 = no wkMedFw1 = age b4 WKModMw1 = none SigDoseWkMw1 = no SigDoseWkFw1 = no SigDoseFw1 = no wkModFw1 = none wkModMw1 = none VactnMedMw2 = age inthobMedMw2 = age inthobMw2 = age PrbfmhmModMw2 = none MainEffProbSocFw2 = age radhlw2 avgcumdosew2 bf4 hmcrModMw2 = none MainEffhmcrMw2 = age wkMedFw2 = age b4 wkMedMw2 = age bf4 MainEffsxlifeMw2 = age bf4 bf40 shjobw2 shrelaw2 radhlw2 MainEffPrbfmhmMw2 = bf4 bf6 bf7 ProbsocMedFw2 = age bf4 radhlw2 hmcareMedFw2 = age bf4 WkhmcrMw2 = age b4 MainEffhmcrw2 = age hmcrModFw2 = none SigDoseHmcrFw2 = yes NumhmcrModMw2 = none SigDosehmcrMw2 = no SigdosehmcrFw2 = yes hmcrMedMw2 = age ageXillw2 SigDosehmcrFw2 = no MainEffhmcareMw2 = age WkMedMw2 = age ageXillw2 wkMedFw3 = radhlw3 age ageXillw3 bf40 bf4m bf1 VactnMedFw2 = age illw2 radhlw2 VacatnModFw2 = none MainEffVactnFw2 = age radhlw2 bf7m SigDoseVactnFw2 = no VactnModMw2 = none vactnModMw2 = none SigDoseVactnMw2 = no inthobMedFw2 = age bf4 illw2 bf4m InthbModFw2 = none MainEffInthbFw2 = age radhlw2 bf4 SigdoseInthbFw2 = no InthbModMw2 = none MainEffInthbMw2 = age radhlw2 shfamw2 SigDoseInthbMw2 = no MainEffMw2 = radhlw2 bf4 bf40 SigdoseMEinthob = no sxlifeMedFw2 = age illw2 radhlw2 bf4 bf4m SxLifeModFw2 = no sxlifeModFw2 = none sxlifeModMw2 = none SigDosesxlifeMw2 = no PrbfmhmMedFw2 = age bf4 PrbfmhmMedMw2 = age PrbfmhmModFw2 = none MainEffPrbfmhmFw2 = age bf4 bf40 SigDosePrbfmhmFw2 = no PrbfmhmModw2 = none SigDosePrbfmhmMw2 = no SigDosePrbfhmMw2 = no MainEffPrbfhmMw2 = bf4 bf6 bf7 ProbsocMedMw2 = age ProbsocModFw2 = none SigDoseProbsocFw2 = yes ProbSocModMw2 = none SigDoseProbsocMw2 = no PrbsocModMw2 = none SigdoseMw2 = none hmcareMedMw2 = age hmcareModFw2 = none MainEffhmcarew2 = age SigdoseHmcareFw2 = no hmcareModMw2 = none SigDoseHmcareMw2 = no NameMedMw2 = age ageXillw2 NumModMw2 = none SigDosehmcareMw2 = no SigDoseWKMw2 = no WkMedFw2 = age bf4 WkModFw2 = none WKModMw2 = none SigDoseWkMw2 = no SigDoseWkFw2 = no SigDoseFw2 = no wkModFw2 = none wkModMw2 = none VactnMedFw3 = age illw3 radhlw3 VactnMedMw3 = age illw3 VacatnModFw3 = none MainEffVactnFw3 = age radhlw3 deaw3 SigDoseVactnFw3 = no vactnModMw3 = none MainEffVactnMw3 = age bf7m radhlw3 SigDoseVactnMw3 = no sxLifeMedFw3 = age bf4 bf4m sxLifeMedMw3 = age illw3 InthbModFw3 = none MainEffInthbFw3 = age radhlw3 bf4 SigdoseInthbFw3 = no InthbMw3 = none MainEffInthbMw3 = age radhlw3 shfamw3 SigDoseInthbMw3 = no sxlifeMedFw3 = age illw3 radhlw3 bf4 bf4m sxlifeMedMw3 = age illw3 sxlifeModFw3 = none MainEffsxlifeFw3 = age radhlw3 bf4 bf4m shrelaw3 shfamw3 SigDoseSxlifeFw3 = no sxlifeModMw3 = none SigDosesxlifeMw3 = no MainEffsxlifeMw3 = age bf4 illw3 radhlw3 PrbfmhmMedFw3 = age bf4 PrbfmhmMedMw3 = age PrbfmhmModFw3 = none MainEffPrbfmhmFw3 = age bf4 bf40 SigDosePrbfmhmFw3 = no PrbfmhmModw3 = none SigDosePrbfmhmMw3 = no SigDosePrbfhmMw3 = no MainEffPrbfhmMw3 = bf1 bf4 dvcew3 bf7m ProbsocMedFw3 = age radhlw3 ProbsocMedMw3 = age ProbsocModFw3 = none MainEffProbSocFw3 = age radhlw3 illw3 Shrelaw3 avgcumodsew3 SigDoseProbsocFw3 = yes ProbSocModMw3 = none SigDoseProbsocMw3 = no MainEffPrbsocMw3 = age radhlw3 shjobw3 hmcareMedFw3 = age illw3 hmcareMedMw3 = age illw3 hmcareModFw3 = none SigdoseHmcareFw3 = no hmcareModMw3 = none MainEffhmcareMw3 = none SigDoseHmcareMw3 = no wkMedMw3 = bf8 age illw3 ageXillw3 wkModFw3 = none wkModMw3 = none MainEffwkFw3 = age MainEffwkMw3 = workM: age bf8 illw3 shjobw3 SigDoseWKMw3 = no SigDoseWkFw3 = no 1421 . pwd /Users/robertyaffee/Documents/data/research/chwk/phase3/Htests/h1tests/h1pt2 1422 . di c(filename) chwide16june2012.dta 1423 . 1424 . sjlog close, replace