Co-Designing a Participatory Community Mapping Method for Informal Sheltering in Puerto Rico

Publication Date: 2021

Executive Summary

Overview

A deeper understanding of strategies to improve informal emergency shelter planning and management can be achieved through a participatory action and design research framework. We explored a novel approach of participatory community mapping (or crowdsourcing) in Puerto Rico to be applied under the physical distancing constraints imposed by the COVID-19 pandemic. Recognizing the resources and capacity limitations of formal disaster management systems, individuals and community-based organizations have filled a vital role in crisis response even before the pandemic. Puerto Rico faces unique risks from climate change, earthquakes, the COVID-19 global pandemic, and the slow recovery from Hurricane Maria. In the face of these cascading disasters, building capacity for community-led disaster planning and management is paramount. Building on previous disaster research and community engagement projects, we collaborated with residents, community-based organizations (CBOs), and humanitarian aid agencies (thereafter agencies) to co-design the process of gathering data for our study using the cultural probe method to crowdsource information from remote and underserved communities while following the limitations imposed by the COVID-19 pandemic. The cultural probe is a unique ethnographic research method which uses multiple methods to collect qualitative data on the lived experience of research participants.

Research questions

Key research questions include:

• How have individuals and CBOs been planning and managing emergency shelters, if at all? What challenges do they face in these efforts?
• What are the information needs of CBOs and their constituencies (including vulnerable populations) related to informal sheltering in the study areas?
• Can rich data from cultural probe techniques, including a community mapping exercise, be collected using during a pandemic scenario?
• Can sufficient information be gathered to plan for the location, resources, and management of informal emergency shelters through these approaches?

With this project, we seek to identify innovative, effective community mapping approaches, especially for at-risk populations. We also hope this work will improve shelter planning and management, strengthen social connections by increasing community participation, and facilitate collaboration and coordination among stakeholders to achieve disaster management goals.

Research Design

We conducted a series of remote participatory action and design research activities, including key informant interviews, co-design workshops, and pilot tests with our civic partners. This helped identify and test key design opportunities for using the cultural probe method to crowdsource local knowledge for shelter planning and management.

Results

Interviews with CBO representatives showed a strong desire to participate in a collaborative planning process that includes the community members, the government, and the CBOs themselves. A key distinction made between formal and informal shelters depended on the lead-time before a disaster arrived. CBOs see these informal shelters as vital to their community and are not necessarily imagined as a brick-and-mortar structure, but rather as a community hub for sharing and delivering material and social resources. The concept of vulnerable populations brings to mind common population subgroups such as the elderly, those experiencing homelessness, and children, but appears to overlook people with medical conditions and those of all ages with access and functional needs. Finally, cultural probe methodology was supported as an effective means to gather rich information to enhance disaster planning. Participatory action research was also shown to be an appropriate way to design probes for use in limited resource settings.

Conclusions and Implications

Participatory Action Research methods can be a valuable tool for designing cultural probes and for listening and responding to the voice of a community. CBOs serve an even more valuable role as health educators, trusted communicators, and emerging responders to meet the growing needs of communities after compounding crises, particularly in the wake of a threat to infrastructure safety. The cultural probe provides the ability to see a place through the eyes of the community and within the context of a pandemic. It has been effective even though pandemic restrictions required a hands-off approach to the distribution, completion, and analysis of our research materials.

Introduction

A deeper understanding of strategies to improve informal emergency shelter planning and management can be achieved through a participatory action and design research framework. We explored a novel approach of participatory community mapping (or crowdsourcing) in Puerto Rico that could be applied under the physical distancing constraints imposed by the COVID-19 pandemic. Formal disaster management systems have limited resources and capacity, so individuals and community-based organizations have filled a vital role in crisis response even before the pandemic. Puerto Rico faces unique risks from climate change, earthquakes, the COVID-19 global pandemic, and the slow recovery from Hurricane Maria. In the face of these cascading disasters, building capacity for community-led disaster planning and management is paramount.

With scarce resources during the pandemic, not all households were able to shelter-in-place for extended periods of time. By building community capacity, some services related to sheltering can be led by local citizens and grassroots organizations. The earthquake swarm of 2019–2020 drove residents out of their homes and many found safety and community at informal shelters rather than formal shelters, which were less desirable to many residents after their experiences during Hurricane Maria.

Building on previous disaster research and community engagement projects, we collaborated with residents, community-based organizations (CBOs), and humanitarian aid agencies (thereafter agencies) to co-design the process of gathering data for our study using the cultural probe method to crowdsource information from remote and underserved communities while following the limitations imposed by the COVID-19 pandemic. The cultural probe is a unique ethnographic research method which uses multiple methods to collect qualitative data on the lived experience of research participants.

We conducted a series of remote participatory action and design research activities, including key informant interviews, co-design workshops, and pilot tests with our civic partners. This helped us identify and test key design opportunities for using the cultural probe method to crowdsource local knowledge for shelter planning and management.

With this project, we seek to identify innovative, effective community mapping approaches, especially for at-risk populations. We also hope this work will improve shelter planning and management, strengthen social connections by increasing community participation, and facilitate collaboration and coordination among stakeholders to achieve disaster management goals.

Literature Review

Cascading Disasters and Emergency Shelters

Disasters continue to grow in frequency and severity and all levels of government are grappling with the limitations of their resources and capacities (FEMA, 2018¹). The emergency management sector is facing increased pressure to direct responses for co-occurring emergencies on top of complex recovery processes in Puerto Rico. Changing demographic characteristics and technology usage trends make the effects of disasters more complex to manage and engaging and empowering individuals and communities to become a more integral part of disaster management has become undeniably vital (FEMA, 2011²; Hore et al., 2020³). One of the key areas that will benefit from improved community participation in places like Puerto Rico is emergency shelter planning and management.

In accordance with FEMA’s National Response Framework, emergency shelters are planned under Emergency Support Function #6 (ESF-6): Mass Care, Emergency Assistance, Temporary Housing, and Human Services (U.S. Department of Homeland Security, 2019⁴). These services are typically coordinated through state and local emergency management agencies. Often the main responsibilities fall to departments of housing or social services. These agencies take the lead on running and staffing emergency shelters. However, there are various reasons why these shelters are not fully used: distance or transportation limitations, lack of trust in government, inadequate conditions, or people feeling safer in their own communities (Lee & Chen, 2018⁵).

The decentralized nature of emergency management structures in Puerto Rico resulted in disparate management styles, structures, and resources for emergency sheltering. These factors, along with prior lived disaster experience, has left many residents in non-urban areas to take sheltering into their own hands, as a community. This scenario is not the first of its kind post-disaster. After Hurricane Katrina, many faith-based organizations stepped up to provide sheltering services and have often been formally engaged in shelter planning and operations (Pant et al., 2008⁶). Non-profit or Community Based Organizations (CBO) are able to fill resource or service gaps and provide locally appropriate services to a community. Throughout the COVID-19 pandemic, mutual aid organizations have played a critical role in fillinge these service gaps, particularly for hard to reach populations. Sheltering is not typically a community-led operation. Formal responsibility for providing shelter often belongs to social services, housing, and emergency management agencies, or larger non-profits, such as the American Red Cross. If communities desire to take on this responsibility, they may require guidance to do so safely and effectively to meet the many needs of their community members.

The public health principle of harm reduction, typically applied to injection drug use or alcohol consumption, is founded on an acknowledgement that a population may engage in an activity that is considered risky so it is in the best interest of public health to create the opportunity for these activities to be carried out as safely as possible (e.g., clean needle exchange programs, condom distribution, etc.) (Ezard, 2001⁷). We can apply this principle to emergent sheltering, where during a disaster, communities establish informal shelters to fill a service gap in their locality. It is critical to acknowledge the phenomena of informal shelters emerging when formal options cannot fully meet the needs of a community and therefore a locally coordinated approach is necessary.(Schlegelmilch et al., 2020⁸). This project aims to ensure these informal shelters are not only located appropriately, but that they also have adequate resources to meet the needs of children and other at-risk populations such as the elderly, those with special health care needs, and people with access and functional needs. The idea is that if we are able to take care of the most vulnerable in a community, particularly children, then it will ultimately build the resilience of the whole community.

Community Information and Participatory Research in Disaster Response

Disaster research and accounts from various studies demonstrate the adaptability, innovation, and responsiveness of community groups in the face of crisis where strong social networks and structures are important factors in recovery after disasters (Aldrich & Meyer, 2015⁹; Liboiron, 2015¹⁰; Whittaker et al., 2015¹¹). However, individuals and groups working outside of the official disaster management system have been undervalued by formal institutional structures and arrangements, which has led to the marginalization of local information and knowledge in disaster risk reduction (Gaillard & Peek 2019¹²). In the context of complex and growing hazards, the understanding of and respect for local knowledge can help agencies improve their disaster management planning; project performance; and acceptance, ownership, and responsibility (Dekens, 2007¹³; Spiekermann et al., 2015¹⁴). These findings call for a different type of disaster information and knowledge management system that values the time, skill, knowledge, and resources that residents and community groups offer (Gaillard et al., 2019¹⁵; Wachtendorf et al., 2017¹⁶). To truly engage and empower individuals and community groups will require the disaster management community to transform their thinking, planning, and practice.

Participatory action research (PAR) seeks to reduce hierarchies between the researcher and participants and focuses research questions on the needs and priorities of communities who are often excluded from decision-making processes. Anecdotally, shelter plans are devised without community input, which creates an opportunity to facilitate such participation. Crucially, knowledge produced by PAR processes are considered valid when it helps to resolve pressing questions in the context of the research location. Thus generalizability is considered to be less important than practicability, and while the results of PAR methods can inform other work in other settings, the knowledge they yield is always understood to be situated and contextual (Greenwood & Levin, 2006¹⁷; Kemmis et al., 2014¹⁸). PAR methods have been paired extensively with design research (Greenbaum & Loi, 2012¹⁹; Hayes, 2011²⁰). There is also increasing interest from disaster researchers to ensure that public participation in disaster planning processes is robust and meaningful. This is accomplished by participation in mapping processes where inequities between experts and at-risk or affected communities can lead to local needs and perspectives being left out of formal disaster management processes (Gaillard et al., 2015²¹; Hore et al., 2020; Soden & Palen, 2018²²).

The constraints imposed by the COVID-19 pandemic, such as physical distancing and uneven and unequal access to electricity and technology, make many standard approaches to facilitating participation in disaster management (such as in-person workshops, design charrettes, or online crowd-sourcing techniques) impossible at this time. Furthermore, the complexity and context specificity of sheltering decision-making and planning made us concerned that highly constrained methods of data collection, such as surveys via SMS or interactive voice response (IVR), would be note capture the depth of information necessary. To address this problem, we experimented with a design research method known as a cultural probe. Cultural probes are used by design researchers to collect information about the perspectives, daily lives, and desires of diverse participants (Gaver et al., 1999²³; Wyeth & Diercke, 2006²⁴). Cultural probes take many forms, but they generally consist of packets of paper materials that are mailed or delivered to participants. Depending on the probe, participants may have several days to a few weeks to complete and return the materials. Probes can take many forms. Depending on the research goals, participants may be asked to draw pictures, provide written responses to short prompts, or draw maps or take photos of particular aspects of their environment. In some cases, designers ask participants to produce creative or artistic responses (Gaver et al., 2004²⁵). They have been used in diverse design settings, including campus sustainability planning (Davis, 2010²⁶), childhood education (Wyeth & Diercke, 2006), and international development (Wyche, 2020²⁷). There are significant debates within the design community around the evaluation of probe results, particularly if the findings are used as research methods as opposed to inputs into individual design processes (Boehner et al., 2007²⁸). Appendix A presents examples of cultural probes that have been used in other studies. We return to this issue in our section on data analysis. Cultural probes have the ability to capture rich data on the lived experiences of an individual in their environment, so we believe this method provides unique perspectives for understanding how residents perceive the community environment and live their daily lives. It can also help identify the priorities of those who require more specialized services in a shelter.

Methods

Research Questions

In collaboration with civic partners in Puerto Rico, we explored how to effectively deploy cultural probes to enhance informal shelter planning, particular community mapping activities that could enhance informal shelter planning. Building on previous research, community engagements, and existing literature in disaster management research, we use a set of premises to guide our research:

• Effective information and knowledge management is foundational to achieving disaster response and recovery goals.
• Local knowledge is a key component of disaster information and knowledge management, but it is not consistently integrated into formal response mechanisms due to functional, structural, and social barriers.
• Leveraging a community’s existing strength and innovation enhances collaboration between individuals, CBOs, and aid agencies to achieve disaster resilience goals.

In this study, we apply PAR and design research approaches to facilitate conversations on the needs and capacities of individuals and CBOs, strengths of and gaps in existing practices, and opportunities for innovation in designing a new participatory community mapping and data gathering approach. In particular, we aimed to answer the following core research questions that correspond with the three key components of our project:

1. How have individuals and CBOs engaged in the planning and managing of emergency shelters, if at all? What challenges do they face?
2. What are the information needs of CBOs and their constituencies (including vulnerable populations) related to informal sheltering in the study areas?
3. Can rich data from cultural probe techniques, including a community mapping exercise, be collected effectively under pandemic restrictions?
4. Can sufficient information be gathered to plan for the location, resources, and management of informal emergency shelters using these approaches?

These core research questions guided our engagements with our civic partners and residents to co-design and test cultural probe processes for crowdsourcing community information about informal sheltering.

Study Site Description

Located in the northeast Caribbean Sea, Puerto Rico is home to 3.2 million people and exposed to a range of natural hazards, including hurricanes, earthquakes, tsunamis, landslides, subsidence, and flooding (Palm & Hodgson, 1993²⁹). In September 2017, two landfalling Category-5 hurricanes, Irma and Maria, caused catastrophic damage to homes and critical infrastructure in Puerto Rico (Segarra, 2017³⁰). Before the island had recovered from the hurricanes, Puerto Rico suffered a series of earthquakes at the end of 2019 and start of 2020. Then the COVID-19 global pandemic struck in early 2020 (Ayala et al., 2020³¹). These back-to-back disasters exposed the widening gaps in providing mass care and sheltering for the residents who had been affected and exacerbated by a pre-existing housing crisis (Hinojosa & Meléndez, 2018³²). The earthquakes forced residents into the streets out of fear and sometimes necessity. It was reminiscent of their experiences during Hurricane Maria (Acevedo & Gutierrez, 2020³³). Across the island, residents and CBOs provided emergency shelter and mass care (Posada, 2018³⁴).

Data, Methods, and Procedures

This project used participatory action and design research methods to help civic partners in Puerto Rico explore the use of cultural probes to gather community information to plan and manage informal emergency shelters under the constraints imposed by the pandemic. The work was rooted in deep community engagement and done in accordance with public health advisories. We conducted virtual primary research, including key informant interviews, cultural probe co-design workshops, and pilot tests, to identify the needs and capacities of stakeholders, co-create design concepts of the cultural probe package, and evaluate the efficacy of the cultural probes. We carried out the project in five stages, as seen in Figure 1. Activities and timeline summary can be found in Table 1.

Figure 1. Research Activities and Procedures

Table 1. Participatory Action Research Objectives and Activities

Recruiting Participants and Planning Research

With the support of a Puerto Rico-based community liaison and research coordinator, we designed an application process to recruit three partner CBOs. The application included a short written component and interview. It was provided to several CBOs that were refered by experts in Puerto Rico. The selected CBOs represent geographic diversity, demonstrate prior experience working directly with local communities on response and recovery, and serve populations that could benefit from the project.

Each selected partner CBO received a microgrant to support their time, effort, and resources for this study, including the distribution of incentives to the study participants they recruited. Moderate incentives were provided to community members and key informants for their participation in the PAR and design research activities. All research protocols were approved through Columbia University Institutional Review Board (IRB), protocol IRB-AAAT5061. All project staff and CBO partners involved in human subjects research, including recruitment, completed the necessary human subjects research training.

Each of the selected CBOs then recruited eight local residents to participate in the co-design workshops and ten others to complete the cultural probe activity packet. A convenience sample of three target subpopulations were recruited by the CBOs. These populations included caregivers of children, caregivers of the elderly, and the elderly population (65+ years old), all of whom were identified as vulnerable or at-risk by the CBOs. Based on expert referrals, we recruited five humanitarian aid agency representatives who work in functions related to emergency shelter planning and management to complete key informant interviews. Key informant interviews were also conducted with CBO representatives. Throughout the research project, we collaborated with CBO and agency representatives.

Co-Designing Cultural Probes with Partners

After recruiting workshop participants (n=28) and consolidating PAR and design research plans, we conducted key informant interviews (n=7) and facilitated three co-design workshops to inform the probe design. Interviews and co-design workshops were conducted in Spanish and facilitated by Puerto Rican research team members. Two-hour long semi-structured key informant interviews with seven civic partner (CBOs and agencies) representatives gathered in-depth information on their emergency sheltering planning and management processes as well as their relevant information needs. The key informant interview guiding questions are included in Appendix B.

In each community, we worked with CBOs to conduct a virtual two-hour co-design workshop with up to eight residents using the video-conferencing platform Zoom. Throughout these co-design workshops, we used a human-centered design process to guide resident participants to co-create the cultural probe concepts. The general workshop process included discovery and problem definition, ideation and prototyping, concept evaluation, and workshop feedback. The workshops were hosted by one facilitator, one notetaker, and one technology support and time-keeping staff member. The research team went through an iterative process and modified workshop activities based on feedback from previous ones to improve participant experience and research outcomes. The key activities and workshop prompts are included in Appendix C.

Findings from the PAR and design research activities were then used to create the final cultural probe packages to be distributed for pilot testing. Based on workshop recordings and notes, we created virtual sticky notes using Miro, an online collaborative workspace, and grouped them by themes related to shelter planning and disaster management, disaster experience, social network, skills/resources, food, mental health, communication/information, normalcy/fun, structure/design. Each of the five researchers were given seven votes to pick their top choices, and preference was given to ideas generated by workshop participants. Among the top choices, we combined similar concepts and modified them for clarity. The research team then reviewed the modified concepts using the following criteria, each receiving a low, medium, and high rating: ease of use for participants, ease of manufacturing, data analysis complexity, ease of distribution/collection, enjoyability, trust-building/community-focused, prioritizing at-risk populations, meeting existing disaster management mandates/protocols, and workshop participant contribution. Finally, based on an evaluation score, estimation of time required for completing the activities, and coverage of key shelter planning and disaster management aspects, we created a cultural probe package consisting of 11 activities and a demographic questionnaire on that covered a range of shelter planning and disaster management functions. The cultural probe package was translated into and administered in Spanish. The co-created cultural probe (in English) used in this study is included in Appendix D.

Distributing Cultural Probes and Collecting Data

The cultural probes were designed to take five to 10 hours to complete over the course of one week. The CBOs distributed and collected the cultural probe packages that consisted of packets of materials and exercises. Each packet included a paper set of activities, disposable camera, pen, colored pencils, and a pencil sharpener. Derived from the PAR and design research findings, the cultural probe packages provided clear guidance for participants to document a range of information on their activities, perspectives, feelings, and motivations relevant to sheltering and mass care in their communities. The participants were able to carry out these activities in their communities and with minimal interference.

Overall, 30 packets were completed (n=30), 10 respondents per community. Participants included 11 males and 19 females. Participant age groups included 18–30 (n=2), 31–49 (n=13), 50–64 (n=8), and 65+ (n=7). Participants by sub-population included caregivers of children (n=8), caregivers of the elderly (n=9), dual caregivers (n=1), and elderly living alone (n=7).

Analyzing Cultural Probe Results

Data from cultural probes is notoriously difficult to interpret in ways that align with the epistemic commitments of other areas of the social sciences because the approach is rooted in design practice and research (Boehner et al., 2007; Gaver et al., 2004). Gaver, credited for popularizing cultural probes, has noted that they are not meant to deliver “a list of facts” about participants and wondered in 2004 if the data that probes return was “impossible” to evaluate in a manner that would be legible to other disciplines (Gaver et al., 2004). However since the time of writing, cross-traffic between ethnography and design research has increased (Dourish, 2006³⁵; Gunn & Donovan, 2016³⁶; Suchman, 2011³⁷) and probe methodology has been taken up in more practical ways than Gaver initially envisioned (e.g. Celikoglu et al., 2017³⁸; Wyeth & Diercke, 2006). Nevertheless, interpretation of probe data remains inherently subjective and requires significant reflexivity (Wyche, 2020).

Our team—composed of a disaster informaticist, a human-centered designer, community psychologist, anthropologist, and a disaster and public health researcher—examined the probe results in a collaborative fashion. Our specific research concerns related to: (a) participants’ lived experience during previous disasters; (b) facets of the local context or environment that may support (or hinder) sheltering during storms; and (c) relationships between individuals, CBOs, and other institutions that would impact shelter planning. We considered these sensitizing questions for the process. After reviewing the data individually, we met twice as a group to discuss emergent themes and share representative examples to allow opportunity for further reflection on our analyses. These discussions helped us examine our own assumptions and biases. This was important given each member of the research team’s different relationships to, and levels of familiarity with, our study site. At the second meeting we agreed upon an initial set of themes to orient our analysis. Several members of the team drafted thematic memos, which were reviewed and edited by the full team.

Disseminating and Evaluating Results

After data analysis, we disseminated the results to all participants and conducted review sessions with CBO and agency partners to evaluate the effectiveness of the cultural probes for meeting the partners’ needs concerning emergency shelter planning and management. Participants in the feedback workshops at the end of the process were given short surveys to complete in order to help evaluate the value of the findings produced for shelter planning and reflections on the process itself. These findings were summarized by one member of the team along with qualitative data collected during the workshops.

Researcher Positionality, Reciprocity, and Other Ethical Considerations

All research activities were conducted after approval of the Columbia University Institutional Review Board. The data and information collected do not contain personally identifiable information or protected health information. In light of the COVID-19 global pandemic, all research activities were conducted virtually to protect the research team, local collaborators, and research participants. The study design was based on principles of community-based participatory research and, as such, directly involves the and collaborates beneficiary communities in all phases of research. Study findings and final product materials were made available to participants and their communities. All interactions with CBOs and study participants were facilitated in Spanish by a Puerto Rican project community liaison and research coordinator based in Puerto Rico. At least one co-investigator was present at all data collection activities.

Findings

Results

Interviews with CBO representatives show a strong desire to participate in a collaborative planning process that also includes the community members and government agencies. We made an important distinction in our research between formal and informal shelters. Informal shelters tend to be used more often when there is not much lead-time before a disaster arrives. Hurricanes provide much more time to prepare than a no-notice disaster such as an earthquake, which may lead more displaced residents to use informal shelters. CBOs see these informal shelters as vital to their communities and are not necessarily imagined as purely brick-and-mortar structures but rather as a community hub for sharing and delivering social and material resources. Communities also need reliable communication networks to receive accurate and up-to-date information and to communicate the community’s needs. Some agencies understand that poor communication does not foster trust with their constituents. The concept of "vulnerable populations” brings to mind the elderly, the homeless, and children, but appears to overlook the medically fragile and those of all ages with access and functional needs. Finally, we found that cultural probe methodology was an effective means to gather information that could be used to enhance disaster planning. The participatory action research approach was an appropriate way to design probes for use in limited resource settings.

“The role of the community [in the planning of an official shelter] is practically none." -CBO Representative

Below, we present initial results from three select probe activities: shelter design, community mapping, and sources of information. Appendix D contains detailed activity instructions.

Shelter Design (Activity 11)

Participants were asked to imagine a hypothetical situation in which relocating to a shelter would be the safest option and then design their ideal shelter. They listed and described the most important elements and services, and made drawings of the shelter considering comfort, safety, and necessary elements. We analyzed the lists and drawings, paying attention to what elements were included, how they were prioritized, the layout and sizing of items, and other details in order to understand the participants’ perception of ideal shelters.

Although a few of the participants focused solely on basic necessities such as bathrooms, beds, and food, many of them enhanced the shelter with options for entertainment, collective spaces and activities, accessibility for people with disabilities or special healthcare needs, and educational and cultural components. Some participants emphasized the need for psychosocial support and community activities to buffer the emotional impact and trauma that can result from experiencing a disaster. Others focused on important characteristics related to the context of natural disasters, including weather-resistant structures and reliable water and energy sources. A few participants considered the unique context of the COVID-19 pandemic, including hand-washing stations, special medical equipment, and distancing between beds.

Community Mapping (Activity 10)

To help identify potential locations for shelters, participants were asked to mark, draw, and label the following elements on a blank page (or pages as needed): their home, place(s) they would put a shelter, their route from home to shelter and describe how they would get there, community landmarks or points of interest (schools, government buildings, etc.), areas they would avoid during a disaster, and areas they think are safe during a disaster. The maps were analyzed with a focus on topography, community features, areas to avoid, and our evaluation of how useful the maps are for shelter planning.

Participants who live close to water commonly identified rivers and beaches as topographical features and areas to avoid. Other areas identified as places to avoid included broken infrastructure and areas prone to landslide. The most common community features participants identified included houses (their own and of neighbors and families), businesses (such as local stores known as colmados), streets and roads, churches, parks and plazas, sports fields (canchas), schools, parking, and community centers. While some participants identified empty lots as potential shelter locations, many considered churches, schools, and community centers to be ideal shelter locations. Most participants identified key community features around their residence and potential locations for shelters. It’s important to note that many of the maps may need additional references, scale, and labeling of streets and landmarks to make them understandable to people from outside their communities.

Trusted Information Sources (Activity 7)

To understand how people obtain trusted information, we asked participants to identify up to four sources of information. We also asked when and how often they go to that source, and if it would be a trustworthy source of information during an emergency.

Among the sources reported, neighbors, family, friends, and other personal contacts (especially those with connections to the government, emergency management and first response, and/or in other leadership positions such as community leaders) were the most common. The other important sources also include radio, television, newspaper, municipality and emergency management entities, social media (such as Facebook).

Demographic Survey

Several questions were posed to respondents regarding the prior shelter experience and anticipated use. Regarding prior shelter use, 73% (n=22) of respondents have never stayed in a shelter, 17% (n=5) have stayed in a formal government-run shelter, and 10% (n=3) have used an informal shelter. In the event of a major disaster 40% (n=12) would opt to shelter-in-place at their current home, 40% (n=12) would stay with a friend or relative, 16% (n=5) would go to a formal shelter, and 3% (n=1) would stay at an informal shelter. Regarding the future use of shelters, 40% (n=12) said they would prefer to stay in their home, 40% (n=12) stated they would stay with a friend or relative, 17% (n=5) would go to a formal shelter, and 3% (n=1) would use an informal shelter.

"[We need] to be recognized in that we can, that we have the resources, and that we are available [to be the] first responders.” -CBO Representative

Discussion of Findings

Qualitative interviews with community-based organizations (CBOs) revealed their essential but overlooked and unnurtured capacity to provide shelter-related services during a crisis. They expressed their desire to be integrated into formal government emergency shelter planning and response efforts and identified their role as a critical link to the needs of their communities regardless of an inherent distrust based on previous disaster experiences. As expected, some CBOs have been assuming responsibility and defining their roles to support community emergency planning since the 2019/2020 earthquakes and Hurricane Maria. These findings suggest the role of these organizations as central actors in a community’s resilience to disasters, and that they have a desire to gain capacity, education, and resources to reach the most vulnerable people in a community.

Findings from the co-design workshops and from the cultural probe data show that participants expressed varied ideas on shelter design. Most striking however, was how many participants considered aspects such as community orientation, psychosocial support, accommodations for vulnerable groups such as nursing mothers or the elderly, and even entertainment as essential elements of shelter planning and design. Such features range far beyond rudimentary views of shelters as simple structures to provide temporary refuge of displaced people for physical security and basic needs. While our participants were not emergency managers, or experts in shelter planning, they have experienced disasters before and have important knowledge about their communities’ needs and were therefore able to provide useful feedback to the shelter planning process.

The time and effort needed to evaluate cultural probe results is significant. It also relies on different methods and sensibilities than researchers or planners would draw upon to assess more narrow or quantitative research instruments. CBOs and planners will require some training and resources to use these tools. As described above, our probes were designed provide more interpretive findings that help to expand the design space in which shelter planners work, construct rich understanding of participants’ lived experience of thorny and personal concepts like safety, shelter, and danger. It also can illuminate surface issues—such as the usefulness of providing entertainment or culturally appropriate food—or other ideas that may be missed through more narrowly targeted research approaches.

This approach was unfamiliar to some members of our research team and collaborators, and so required additional attentiveness and care in both evaluating and communicating the results. Further development of the cultural probe may provide response organizations with a better way to understand the desires of the people who may be in the extreme circumstance of needing shelter when no safer alternative exists. This process will not only help plan for more contextually appropriate shelters but offer an opportunity to build trust between all stakeholders.

Conclusions

Key Findings

Participatory action research methods to design and deploy cultural probes as a method of community inquiry can be a valuable tool for listening and actively responding to a community. CBOs serve an even more valuable role as health educators, trusted communicators, and emergency responders to meet the growing needs of communities during compounding crises, particularly in the wake of a threat to infrastructure safety.

Implications for Public Health Practice

Climate change is a public health crisis which is continuing to stress the resources and human capital of disaster response systems. In 2015, an estimated 67% of Americans believed climate change was making disasters more severe, and 40% of Americans were not confident in the government’s ability to adequately respond to a major disaster (Petkova et al., 2016³⁹). One year after Hurricane Maria, many people in Puerto Rico were still struggling to meet their basic needs. According to a poll, Puerto Ricans saw a failure in the government response to Hurricane Maria at all levels including municipal governments, the Puerto Rican government, and federal agencies, giving negative ratings to President Trump and former Governor Ricardo Rosselló (DiJulio & Muñana, 2018⁴⁰) for their handling of the disaster. There were major concerns about whether authorities would be able to help if another storm hit the island. The majority of Puerto Ricans worried that the government was still not prepared to deal with future hurricanes and believed that most residents were not ready.

Emergency shelters, also known as mass care facilities, provide immediate shelter after a disaster when a person’s home or residence is not livable. These temporary shelters provide the community with food, social services, the ability to apply for disaster aid, access disaster case workers, and first aid or medical referrals for specialty care. These shelters are critical to a community’s resilience because they buffer the impact of disasters on individuals and households. A well-run shelter that addresses the specialized needs of vulnerable populations requires significant planning and serves a critical public health function—but it is important to note that recovery and response cycles frequently overlap and disaster plans have fallen short of community expectations. Disaster systems, volunteers, formal disaster response agencies, and humanitarian aid organizations in Puerto Rico have been fatigued by multiple disasters.

Community-based participatory research methods connects directly with the value of social capital before and after a disaster (Aldrich & Meyer, 2015; Nakagawa & Shaw, 2004⁴¹). By involving communities in shelter planning and design we hope to provide an example of a mutually beneficial activity that strengthens the fabric of a community and provides a vital service in a crisis. This process helps locate and identify the most marginalized and at-risk members of a community—including the elderly, medically fragile, people with disabilities, access, and functional needs—who require the assistance of the whole community. This method of data collection may be less familiar to the public health research community, but it could hold promise as a tool for gaining a deeper contextual understanding of affected communities. The role of place in a person’s health and well-being is deeply connected to their community—geographic space, cultural fabric, and connectedness (i.e. social capital). The cultural probe provides researchers and policymakers with the ability to see place through the eyes of the community. It was also effective while pandemic restrictions, such as social distancing, were in place, given the hands-off distribution, completion, and analysis.

Finally, we worked to facilitate collaboration between emerging citizen groups and official disaster management agencies to bridge the planning and communication gap. Organizational, functional, and cultural differences can make collaboration and communication between key stakeholders challenging. By facilitating partnership building, this project bridges the current gaps among these groups using new ways of information management and sharing.

Dissemination of Findings

The project leads will present all findings and research materials to collaborating community members in a webinar no later than the project end date of October 30, 2021. During this event, we will share the results of the probes with anyone who participated in either the workshops or the cultural probes. The webinar will walk viewers through the participatory action research, design process, and explains the implications of the findings to community organizations and activists in Puerto Rico and other territories. In addition to this report, the team will draft a paper for peer-reviewed publication to present the novel concepts we have tested and their potential for implementation on a practical level.

Limitations

The exploratory nature of our research serves as a proof of concept, but the design and implementation processes require further analysis. The scalability of this research process, given the resource limitations that CBOs often face, and the constraints of working during a global pandemic are a major concern, but can be addressed with a well-matched research partner and modest funding. These findings may not be generalizable to other communities given its small sample size and how new the research methods are. Furthermore, these results are also not generalizable to all vulnerable populations in Puerto Rico. These limitations may be due to recruitment strategies, the fact that the interactions took place online, or the topography of the selected study areas.

Future Research Directions

Further research is needed to validate the scalability of using cultural probes as a method for integrating community knowledge and information for informal sheltering or other public health issues. Further refinement of this specific probe would be recommended with a more robust examination of how actionable it is for policymakers, emergency planners, and community-based organizations. Implementation science can be used to examine process and feasibility issues. We recommend further exploration of the shelter as a fluid concept using a cultural anthropological approach to determine if alternative disaster shelter models would be used and trusted more.

References