Does the Public Ignore Female Governors on the COVID-19 Pandemic?

Publication Date: 2022

Abstract

Preparedness during public health crises depends heavily on the effective communication of public leaders. Effective communication and leadership outcomes are also shaped by socially constructed gender role expectations that attribute instrumental leadership qualities to men. The current COVID-19 outbreak offers a natural experiment setting to study if women in key leadership roles influence public perception of leader effectiveness and willingness to take protective actions during emergencies in ways similar to their male counterparts. Although the relatively new, COVID-19 literature has started to examine these issues, but its findings are mixed. Furthermore, it does not examine how female governors are viewed in the context of COVID-19. This study was designed to fill in these gaps in the literature. Its main research questions were: (a) Are residents in male governed states more likely to perceive them as competent leaders compared to residents in states with female governors? (b) Are residents in female governor states more likely to perceive their governors as more caring leaders compared to residents in male governor states? (c) Are residents in male governor states more likely to have engaged in COVID-19 protective behavior in the past than residents in female governor states? (d) Are residents in male governor states more likely to demonstrate protective behavioral intent than residents in female governor states? The study was based on 1,203 surveys conducted in eight U.S. states: Alabama, Iowa, Oregon, Rhode Island, Washington, Connecticut, South Carolina, and Nebraska. It revealed a negative effect of the governor’s gender on resident future protective behavioral intent in addition to a positive influence of residents’ rating of governor leadership competence on taking protective steps. The study’s findings on female residents’ compliance with COVID-19 orders were consistent with the literature.

Introduction

Preparedness in times of public health crises such as the COVID-19 pandemic depends on leaders’ effective communication with the public. Such communication is critical for the public to understand the levels of risk and to take recommended protective actions to mitigate the negative effects of the crisis, such as the number of lives lost, the number of confirmed cases, and the toll on businesses and the economy. Despite a growing literature on effective disaster communication, we have a limited understanding of how the social construction of leadership roles, specifically with respect to gender, shape the public’s perception of leader effectiveness as well as the public’s risk perception and compliance with protective measures in disaster contexts.

The current COVID-19 outbreak offers a natural setting to study these overlooked questions in the disaster literature. Hence, the goal of this project is to advance scientific knowledge on gender, leadership, and disaster communication within the context of COVID-19. To achieve the study’s objective, the study sets two specific aims: (1) to explore differences, if any, in how the public perceives public leaders’ effectiveness during the COVID-19 pandemic; and (2) to examine if there are gender differences in the public’s response to male and female state governors’ leadership and calls for protective actions with regards to the COVID-19 outbreak.

In the next section, we provide a review of the relevant literature on leadership gender bias and emergency management. We then present our methods of data collection and survey findings. Next, we discuss the findings of the study in the broader leadership literature and conclude with the study’s policy implications and future directions for research.

Literature Review

This study links the leadership gender bias literature with the emergency management literature. The leadership gender bias literature suggests that women are vastly underrepresented in key leadership roles across different sectors (Bishu & Headley, 2020¹) and that the gender gap in leadership and positions of authority cannot be explained by gaps in human capital or differences in work performance (Alkadry et al., 2019²; Mitra, 2003³; Smith, 2002⁴). Rather, it can be attributed to challenges women face in accessing leadership roles (Scott & Brown, 2006⁵) and to gendered perceptions of leadership (Heilman, 2004⁶). The think-manager-think-male framework explains the notion that leadership roles are commonly associated with masculine attributes of being authoritative and directive as opposed to feminine attributes of being empathic and caring (Croson & Gneezy, 2009⁷; Sabharawal, 2015⁸; Schein, 1975⁹, 2001¹⁰; Schein & Mueller, 1992¹¹; Vinkenburg et al., 2011¹²). Such socially constructed gender role expectations attribute autonomous and instrumental leadership qualities to men (Eagly & Karau, 2002¹³; Eagly et al.,1992¹⁴; Sabharawal, 2015). Thus, when women occupy male-dominated roles, their autonomy and leadership influence are questioned. Research also shows that when it comes to job performance evaluation of women who occupy male-dominated roles, women experience pro-male evaluation bias (Johnson et al., 2008¹⁵; Tsui & Gutek, 1984¹⁶).

The emergency management literature acknowledges the centrality of effective communication in emergency contexts. Boin and colleagues (2005¹⁷) and Boin and t‘Hart (2003¹⁸) asserted, for instance, that core responsibilities of leadership in emergency management encompass swift decision-making, coordinating implementation of crisis mitigating measures, and effective communication. In the context of the Ebola public health crisis, Yang (2019¹⁹) highlighted the importance of message design in communicating with the public and in reducing the public’s psychological distance to perception of risk. Several studies also focused on the public’s perception of risk and protective action behavior. Some of these studies incorporated the gender dimension of emergency communications. Their focus, however, was on gender differences in perception and reporting of risk (e.g., McDowell et al., 2020²⁰; Olofsson & Rashid, 2011²¹; Riad et al., 1999²²). An increasing body of research at the intersection of emergency management and communication studies examine psychological dimensions of communication in emergency management (Rickard et al., 2016²³; So & Nabi, 2013²⁴; Spence et al., 2012²⁵; Yang, 2019). These studies identify social and spatial distance between leaders delivering risk communication messages and the public as important predictors of community risk perception and risk behavior (Yang, 2019; Zwickle & Wilson, 2014²⁶). No research to date has explored the relationship between socially constructed leadership role perceptions and the public’s risk perception and risk behavior.

There is a growing COVID-19 literature on female leadership. Several studies examined the impact of female and male leadership on public health outcomes in an international context. Coscieme et al. (2020)²⁷, for instance, compared such outcomes in 35 countries and revealed lower COVID-19 deaths per capita and more effective flattening of the disease curve in countries with female leaders. Still others elaborated on how positive media coverage of female leaders challenges our conception of a good leader and traditionally negative social construction of female leaders in the political sphere (Cherneski, 2020²⁸; Johnson & Williams, 2020²⁹). Some even go further and suggest that increasing support for female leaders’ COVID-19 policies could help relieve the pandemic’s disproportionate burden on vulnerable groups, reduce gender inequities, and contribute to overall development (Piazza & Diaz, 2020³⁰).

The literature that indicates better performance of female leaders globally attributes the difference in COVID-19 public health outcomes in female vs. male led countries to several factors. Coscieme and colleagues (2020) argued that women-led countries introduced restrictive measures early in the pandemic; placed more emphasis on public health and social equalities than economic growth; and pursued a more collaborative governance approach (Coscieme et al., 2020). Some studies highlighted the importance of female leaders’ characteristics, such as sincerity, empathy, resilience, and not being afraid to show their vulnerabilities (Cherneski, 2020) and the ability to blend masculine and feminine leadership traits for androgynous leadership (Blake-Beard et al., 2020³¹). Other studies, however, questioned the validity of research findings on female leaders’ pandemic performance. Piscopo (2020³²) argued that countries led by female leaders would have had low COVID-19 mortality rates regardless of leadership gender as they are high-income, liberal democracies with good governance practices.

The literature on female leadership across the United States is also evolving, both on the impact of the governor’s gender on adoption of COVID-19 restrictions and on public compliance with restrictions. Shay (2020³³) found no impact of the governor’s gender on adoption of stay-at-home orders, although states with public health agencies led by females introduced stay-at-home orders earlier than states with male leaders. Through a randomized study, Bauer and colleagues (2020) investigated public compliance with governors’ COVID-19 orders on social distancing, mask wearing, and contact tracing. This study found little impact of the leader’s gender overall on public compliance with COVID-19 policy recommendations. However, public compliance increased slightly when the order came from a co-partisan female leader and was more personally invasive (e.g., contract tracing). These findings contrasted with earlier literature, which suggested that male leaders are viewed as competent in a national security crisis (Holman et al., 2019³⁴).

The two literatures on leadership gender bias and emergency management have not yet examined if women in key leadership roles influence the public’s perception of leader effectiveness and their willingness to take protective actions in similar ways to their male counterparts in emergency contexts. Although the relatively new COVID-19 literature has started examining these issues, its findings have been mixed. Furthermore, it has not examined how female governors are viewed in the context of COVID-19. This study was designed to fill in these gaps in the literature.

After reviewing relevant literature, we develop four research questions that guide our empirical investigation and interpretation of results:

1. Are residents in male governor states more likely to perceive their governors as more competent leaders compared to residents in female governor states?
2. Are residents in female governor states more likely to perceive their governors as more caring leaders compared to residents in male governor states?
3. Are residents in male governor states are more likely to have engaged in COVID-19 protective behavior in the past than residents in female governor states?
4. Are residents in male governor states more likely to demonstrate protective behavioral intent than residents in female governor states?

Methods

Study Context

We conducted the study in eight states in the United States: Alabama, Iowa, Oregon, Rhode Island, Washington, Connecticut, South Carolina, and Nebraska. To determine these states, we first identified four states led by female governors, with two governors from each political party. We then matched each female governor-led state with a male governor-led state using a state similarity index. This index was created based on states’ demographic and socioeconomic characteristics. While considering the states, we also looked into COVID-19 positivity rates at the time of survey administration and where the rate at which COVID-19 spread was of concern to state leadership. Our final selection of states included two states each with conservative female governors (Alabama, Iowa), conservative male governors (South Carolina, Nebraska), liberal female governors (Oregon, Rhode Island), and liberal male governors (Washington, Connecticut) at the time of the study. We collected our survey data between July 16 and August 8, 2020, after the declaration of COVID-19 as an emergency on March 13, 2020, but prior to vaccine development and distribution.

Data and Method

To investigate the four research questions presented earlier, we received approval for the study from the Institutional Review Board (IRB) at the University of Colorado, Denver and implemented a survey study design embedded in a natural setting (Leatherdale, 2019³⁵). The survey instrument included questions on participants’ demographic characteristics (including their political affiliation), COVID-19 risk perception, media exposure, past and future protective behavior, and their assessment of their governor’s leadership performance during the COVID-19 pandemic. The survey was conducted online via the CloudResearch participant sourcing platform in the eight states included in the study.

Sampling and Participants

Study participants were recruited via the CloudResearch platform using quota by gender and race/ethnicity. A total of 1,203 residents participated in the survey. Of the total participants, 47% (N=564) were male and 53% (N=630) were female, with nine residents choosing not to identify their gender. Our sample consisted of 35.65% (N=426) members of minority groups and 64.35% were White participants (N= 769). Survey respondents were approximately equally distributed among the eight participating states.

Dependent Variables

Broadly speaking, this study assessed two aspects of statewide emergency leadership, preparedness, and response: public assessment of public leader performance during COVID-19 emergency response and protective action behavior. Our investigation of residents’ assessment of leaders’ performance considered leader competence and leader care. By doing so, we aimed at investigating both traditional leadership qualities of competence but also recognize the role that nurturing leadership qualities could play during a pandemic in shaping residents’ behavioral actions related to COVID-19. Thus, based on the existing literature, we created additive index variables for both leader competence and leader care variables from a series of survey questions that measure leader competence and leader care. Questions that measured leader competence had a reliability score of .89 and those that measured leader care had a reliability score of .94. Finally, our last outcome measure assessed three dimensions of the public’s protective behavior by considering (1) individual-level behavioral adjustments; (2) environmental-level behavioral adjustments; and (3) social-level behavioral adjustments. We also asked residents about their past COVID-19 protective behavior and their future intent to engage in protective behavior. For protective behavior, we used three broad categories recommended by the World Health Organization (2020)³⁶: personal protective measures (e.g., wearing a mask), environmental measures (e.g., staying at home), and social measures (e.g., social distancing from friends and relatives). Again, we created an additive index outcome measure that represents residents’ past behavior modification and their future behavior modification intent. Questions that measured residents’ past COVID-19 protective behavior had a reliability score of .87 and those questions that measured residents’ future protective behavior intent had a reliability score of .89. Table 1 presents descriptive statistics of the dependent variables included in the regression analysis.

Table 1. Descriptive Statistics for Leader Competence, Leader Care and Resident Protective Behavior Dependent Variables

Independent and Control Variables

There were two key independent variables of interest included in this study. First, we were interested in evaluating the effects of governor gender across all four outcome variables included in our study. We were also interested in assessing the interactive effects of governor gender with governor political party affiliation on all four outcome variables included in our models. We controlled for a list of resident characteristics: gender, political affiliation, race, level of education, household income, family member with underlying medical condition, follow governor press release, follow current news, follow COVID-19 news, and perception of COVID-19 as a threat. Table 2 summarizes descriptive statistics of the control variables included in our analysis.

Table 2. Descriptive Statistics for Control Variables

Data Analysis

We used an ordered logit regression model to quantitatively estimate the effects of leader gender on residents’ assessment of leader performance. The first two dependent variables of interest estimated the effects of governor gender on residents’ perception of governor leadership performance and the other two dependent variables estimated the effects of governor gender on residents’ past behavioral action and future intent to modify behavior related to COVID-19. Additionally, we observed the main and interactive effects of governor gender along with governor political party affiliation.

Preliminary Findings

We ran ordered logit regressions models to explain the impact of leader gender and leader political party affiliation on the outcomes of interest. Table 3 includes results from ordered logit regressions of governor gender and governor political party affiliation on residents’ evaluation of governor leadership qualities, governor competence (Model 1) and governor care (Model 2), residents’ past protective behavior (Model 3), and residents’ future protective behavioral intent (Model 4). In all models, we assessed the effects of governor gender and political party affiliation on residents’ assessment of leader qualities and their behavioral response to COVID-19. We also observed interactive effects of the two conditions on the outcomes of interest.

Our analysis demonstrated that governor gender shows a main effect only on residents’ future protective behavior intent but not on the other three outcome variables in our models. Governor political ideology, on the other hand, has the strongest effect on residents’ rating of governor leadership, competence, and care. Being a female governor decreases the odds of resident intent to engage in protective behavior in the future. Model 1 and Model 2 also show that residents in conservative governor states have lower odds of rating governors as competent and caring. Despite governor gender not having a significant effect on residents’ rating of governor leadership qualities, the interactive effect of governor gender and political party affiliation appear to have a positive and statistically significant effect on residents’ assessment of governor leadership competence. We observe that being a female and conservative governor increases the odds of residents’ rating governors higher on leader competence and leader care. However, the positive effects that we observed from the interactive effect of governor gender and political party affiliation did not yield a similar outcome when considering residents’ past and future protective behaviors.

Reflecting on some of the control variables included in our models, we noted that being a female versus a male resident increased the odds of residents rating governors higher on leader competence as well as the odds of residents rating higher engagement in past and future protective behavior. Being a resident with family members that have underlying conditions decreases the odds of rating a governor high on leader competence and leader care. We also observed that residents that follow governors’ press release related to COVID-19 pandemic have higher odds of rating governors high on leader competence and leader care. Furthermore, we find that residents that follow governors’ press releases have higher odds of engaging in past and future protective measure behaviors. In addition, higher rates of perceiving COVID-19 as a threat are associated with higher odds of residents rating governors high on both leader competence and leader care, and they are more likely to engage in protective measure behaviors. Lastly, we find that residents rating governors highly in leader competence increases the odds of residents engaging in past and future prospective measure behaviors, thus demonstrating positive effects of residents’ perception of leader competence on behavioral response to COVID-19. Yet, we observe no effect of residents’ perception of leader care on COVID-19 behavioral response.

Table 3. Ordered Logit Results by Governor Gender and Governor Political Party Affiliation

Discussion

There are four key takeaways from our preliminary findings. First, we failed to find evidence of a main effect of gender on residents’ assessment of governor competence and care. This finding is similar to findings from earlier studies that indicated little to no effect of a governor’s gender, albeit in the context of adoption of COVID-19 policies (Shay, 2020). Similar to Bauer and colleagues’ (2020³⁷) study, there were no effects of gender on residents’ past COVID-19 protective behavior. We only observe a negative effect of governor gender on residents’ future protective behavioral intent, which could be explained by gender resentment that decreases compliance with government policy and recommendations (Kalaf-Hughes & Leiter, 2020³⁸). Second, and most importantly, we found a stronger effect of the interaction of gender and governor political party affiliation on residents’ assessment of governor leadership qualities of competence and care. We find that female conservative governors are rated higher on the two leader assessment measures. However, despite residents’ high rating of governors on leadership qualities, it did not appear to have a similar effect on residents’ past protective behaviors. On the contrary, we observe that residents in female conservative governor states are less likely to have engaged in COVID-19 protective behaviors in the past, and we find no statistical difference in their intention to engage in protective behavior in the future. We attribute these findings in part to the politicization of the COVID-19 pandemic and mixed messaging for protective behavior (Grossman et al., 2020³⁹; Halpern, 2020⁴⁰; Hart et al., 2020⁴¹). Third, we find that compared to male residents, female residents were more likely to rate their state governors higher on leader competence, and they were more likely to engage in both past and future COVID-19 protective behaviors. Our results suggest that gender plays a role in how residents evaluated leaders and responded to calls for protective action during the pandemic. These results are in line with prior studies that indicate higher compliance with public health orders among female residents in the United States and elsewhere (Clark et al., 2020⁴²; Czeisler et al., 2020⁴³; Khubchandani et al., 2020⁴⁴). Lastly, we found a positive relationship between residents’ rating of governor leadership competence and their likelihood of engaging in protective behavior. This could be indicative of the fact that residents’ perception of traditional leader performance is an important factor associated with behavioral response to the COVID-19 pandemic.

Conclusion

In this study, we linked leadership gender bias literature with emergency management literature by studying how the public’s perception of leader effectiveness and intention to engage in protective behavior in the future change based on governors’ gender in the United States. Based on a survey conducted in eight states, the study revealed a negative effect of the governor’s gender on respondents’ future protective behavioral intent in addition to positive influence of residents’ rating of governor leadership competence on taking protective steps. The study’s findings on female residents’ compliance with COVID-19 orders were consistent with the literature.

The study offers several directions for future research. First, this study was limited to eight states in the United States. Similar studies could be conducted in other states and in other contexts for a more comparative perspective. Second, there is a need to conduct in-depth studies in conservative and liberal states to better understand the factors influencing why some leaders are perceived to be more competent and caring than others. Third, additional studies could be conducted to further investigate the differences between past and current protective actions between female and male residents. Fourth, future studies could examine the interactions between the governor’s gender and the resident’s gender as well as between the governor’s political affiliation and the resident’s political affiliation. Fifth, the content of governors’ executive orders could be analyzed, as in the case of Curley and colleagues (2021⁴⁵) to see if there are differences in how female and male governors communicate (e.g., in communal or agentic terms).

References