Innovative Participatory Bilingual Data Analysis with Latinx/@ Immigrants: Language, Power, and Transformation

Abstract

Objectives:

The insights of Latinx/@ immigrants are essential to developing interventions that better address complex multi-level phenomena impacting mental health. Despite important advances in methods that genuinely embody participatory research practices, attention to collaborative data collection, analysis and dissemination are limited. Our aim is to describe the development and implementation of research practices to address these gaps through an emphasis on and understanding of the centrality of language in collaborative research processes.

Methods:

Guided from the outset by community-based participatory research (CBPR) principles, our community-academic research partnership recognized the importance of developing and intentionally studying our collaborative processes. As part of an ethnographic interview study with 24 Latinx/@ immigrants, a community-university research team developed innovative methods, including practices related to research team meetings, data collection, analysis, and dissemination, which we documented through ongoing discussion and reflection.

Results:

The resulting participatory research processes were grounded in a theoretical framework of praxis and language and included six innovative and iterative stages: 1) establishing the research team, 2) planning the interview process/data collection, 3) developing the data analysis methodology, 4) interpreting findings to adapt the intervention, 5) integrating results of the participatory process into the analysis, and 6) data analysis for dissemination.

Conclusions:

A focus on praxis and language revealed how the language of research structures power, meaning, feeling, collaboration, analysis, and transformation. We also found that bilingual participatory analytic processes have important implications with respect to achieving genuine inclusion in rigorous research that moves towards equity for Latinx/@ immigrants and other populations.

A foundational narrative of the United States is that it is a country of immigrants who have been essential to the establishment, growth, and identity of the nation-state. In 2017, the United States foreign-born population reached 44.4 million (Radford, 2019). Latinx/@¹ immigrants will continue to have a strong demographic presence for the foreseeable future (Radford, 2019). A growing understanding of the health disparities that many immigrant populations experience and their rootedness in social inequities is a more recent phenomenon. To understand how complex sociopolitical contexts shape migratory and integration experiences and immigrant health, community-based participatory research approaches are essential. Community participation in research strengthens the ability of research to address inequity and positively impact communities (Wallerstein & Duran, 2010). Historically research has been conducted on or about rather than with immigrant communities (Vaughn et al., 2017). When considering how to improve the health and well-being of immigrant populations, it is important to center the perspectives of immigrants themselves. The research reported on here describes innovative collaborative data analysis processes that serve to advance current participatory practices and methodology through a detailed explication of staged, incremental methods. These processes are grounded in participants’ language and experiences, which results in research that is more genuinely participatory and that produces more robust and valid data (Fox et al., 2019). Thus, this work contributes to efforts to advance the science of reducing social inequities and health disparities through bridging theory and practice.

Latinx/@ Immigrant Mental Health

Latinx/@ immigrants’ health outcomes are often viewed within the Hispanic health paradox, with numerous studies documenting that many health outcomes for recent immigrants are better than those for U.S.-born counterparts (Franzini et al., 2001; Ruiz et al., 2013). Thus, they are frequently overlooked in mental health research and development of appropriate interventions (Miranda et al., 2003), despite mounting evidence of mental health inequities and disproportionate exposure to trauma (Alegría et al., 2008; Breslau et al., 2006; Fortuna et al., 2008). The impacts of trauma exposure throughout the migration process and post-resettlement stressors put Latinx/@ immigrant families living in the U.S. at increased risk for experiencing depression, anxiety, and post-traumatic stress symptoms (Bridges et al., 2010; Familiar et al., 2010; Salas et al., 2013). The recent social, legal, and economic context of uncertainty, discrimination, stigma, lack of access to resources, fear of deportation and resulting family separation has also had a critical impact on Latinx/@ immigrants’ mental health in the U.S. (Ayón & Becerra, 2013; Garcini et al., 2017; Sangalang et al. 2019; Sullivan & Rehm, 2005).

Although a growing body of research points to the need to address socio-structural determinants, most available mental health interventions do not adequately attend to structural factors and were not designed, developed or adapted for Latinx/@ immigrants (Cervantes et al., 2019; Garcini et al., 2017; Pérez et al., 2008). The lack of these kinds of mental health interventions is related to the fact that research studies on Latinx/@ immigrants’ health and mental health have traditionally been conducted with data collected from immigrants leading to interventions that are designed for immigrants and research results that are about immigrants (Chang et al., 2013). While CBPR approaches have been used with Latinx immigrants, very little has focused on mental health (Stacciarini et al., 2011; Vaughn et al., 2017).

Community-Based Participatory Research (CBPR)

CBPR has been at the forefront of collaborative efforts to address health inequities (Minkler & Wallerstein, 2008; Wallerstein, Duran, Oetzel & Minkler, 2018; Wallerstein & Duran, 2006). This approach to research provides a framework to work collaboratively with communities impacted by health disparities to research, analyze, and intervene in complex social phenomena, such as the migration experience. CBPR furthermore addresses power inequities between researchers and community members to ensure that research topics, methodologies and outcomes are beneficial to communities. CBPR approaches have yielded important research that has had positive benefits for communities and contributed to a shifting paradigm in public health and social science disciplines that seeks to incorporate the knowledge, lived experiences, and expertise of communities in order to understand and facilitate meaningful collaboration among academic researchers and community members to achieve sustainable change and social justice (Cacari-Stone et al., 2014; Minkler et al., 2012).

Including community members as co-researchers results in more valid research (Vaughn et al., 2017) because the knowledge of community and university research team members is complementary and because valuing community members’ contributions builds trust (Cashman et al., 2008), which results in better understanding of local context and incorporation of community understandings into research (Kerr et al., 2010). Fox, Easpaig and Watson (2019) note that research that moves beyond capturing experiences to shared knowledge production “yields more sophisticated accounts than could be done from the social located perspectives of researchers who often occupy different and more privileged positions than the communities whose interests they seek to promote” (p. 236). Further, addressing differing abilities, social locations, and interpretations within data analytic processes is an important aspect of acknowledging and addressing power dynamics (Foster et al., 2014), which is a foundational goal of CBPR. Such approaches however have been relatively scant in the CBPR literature. Most studies, even recent studies, continue employing a CBPR framework that does not span the entire research process from initiation to dissemination (DiMascio et al., 2020; Torres et al., 2020) and rarely includes community participation in data analysis (Valentín-Cortés et al., 2020).

CBPR and Research Methods

In 2008, Cashman and colleagues reviewed 60 studies that used CBPR approaches, documenting benefits of greater research participation rates, strengthened external validity, and increased community capacity. However, they also noted that the majority of studies included in their review described community involvement primarily in the study design and data collection phases (principally around recruitment), and few described community involvement in data analysis and interpretation of findings. Almost a decade later, Vaughn and colleagues’ (2017) review of CBPR conducted with immigrants also found that very few studies included participants in data analysis. Moreover, they highlighted a lack of description of collaborative processes that makes it difficult to ascertain the level of community involvement and character of partnerships. Although a few studies of collaborative data analysis in CBPR have emerged (Cashman et al., 2008; Flicker & Nixon, 2014; Foster et al., 2014; Rosenthal et al., 2014), examples of participatory research that provide detailed descriptions of the analytic processes and associated outcomes are sparse.

CBPR and Language

CBPR principles include recognition of the importance of language, culture, and power dynamics (Cashman et al., 2008; Chang et al., 2013; Stacciarini et al., 2011; Vaughn et al., 2017; N. Wallerstein & Duran, 2010), and CBPR literature widely reports the incorporation of bilingual approaches, including bilingual survey instruments and interviews, and bilingual staff members (e.g., De La Torre et al., 2013). However, most studies do not discuss language beyond mere translation (Kao et al., 2004). The CBPR literature on conducting research with Indigenous communities is more robust when considering approaches that build on cultural language and knowledge (Kerr et al., 2010; Kovach, 2005; Lavalee, 2009; Loppie, 2007; Wilson, 2008). However, Latinx immigrants’ understandings have not been centered as extensively as decolonizing approaches created by Indigenous researchers and communities. Increasing diversity in the research process is not only about increasing community participation, but also the actual process of co-creating knowledge to achieve social justice and meaningfully impact communities (Parra-Cardona et al., 2020).

Theoretical Framework: Praxis and Language

When considering participatory methods of data analysis, attention to methodology is essential, which is understood to be theorizing of methods that produce legitimate knowledge claims (Fryer & Laing, 2008). Praxis, or attending to what people do (in this case, in research spaces), is a helpful guide when considering a theoretical framework that illuminates methodological considerations. Social theorists concerned with practice theory, including Bourdieu (1977) and Ortner (2006), suggest attending to both structure and agency and the ways these are mutually constituted, which is essential to gaining an understanding of the role of individuals and groups in creating social change. Structure refers to social and institutional shaping and restraint of individual agency, which concerns both conscious motivation and the ability to act.

Language plays an important role because, “The human animal is born into language and comes into being through its terms” (Alexander 1994, p. 277). Carter and Sealey (2000), for example, talk about the importance of language to our psychobiography, or how we see ourselves over time. Related to the issue of structure and agency, a focus on language helps to underscore both the importance of individual action and the socio-structural systemic context that constrains, shapes and enables language. In this way, understanding how power relations infuse individual identity, choice and decision making is critical to how people move in the world (Ortner 2006). Symbolic interactionist theories (Goffman, 1959, 1967; Schwalbe, 1983) argue that language is co-created, thus underscoring the importance of attending to its use in participatory research spaces. Researchers have also recognized the importance of attending to positionality and power in multilingual settings (Giampapa & Lamoureux, 2011). A focus on language also ensures that research methods pay close attention to experiences and processes from participants’ point of view, elaborating meaning from their perspective and building on that in analytical phases. This notion is in line with Eisenhart’s call to align research methods with the social worlds and voices of research participants (2001). This, in turn, ensures that analytical processes are collaborative in nature. As each of these phases build on each other, transformative processes are implemented at the research level.

Within this context of increasing efforts to create research partnerships that produce transformational processes and sustainable interventions to promote well-being and reduce inequities, it is important to explore specific methodological approaches to both research and intervention that can achieve these goals. This article describes the ways in which innovative bilingual collaborative processes were developed and incorporated from the outset of a new research partnership with Latinx/@ immigrant community members and organizations, with a specific focus on the typically neglected phases of data analysis, interpretation, and dissemination in CBPR. Emphasizing genuinely participatory approaches in every step of the process not only generates better research, but also holds the promise of producing the transformative processes on a small scale that we are trying to replicate in society at large.

Methods

Research Setting/Context

In 2017, the Transdisciplinary Research, Equity and Engagement (TREE) Center for Advancing Behavioral Health was established as one of twelve NIMHD-funded Centers of Excellence on Minority Health and Health Disparities (U54 MD004811). The Immigrant Well-being Project (IWP) was funded as one of two core research projects within the TREE Center. In recognition of the increasingly hostile climate and risks for Latinx/@ immigrants in the United States, the goal of the IWP study was to test a transdisciplinary ecological approach to reducing mental health inequities among Latinx/@ immigrants by adapting a multilevel social justice intervention (Refugee Well-being Project; or RWP) designed to address post-migration socio-structural stressors of refugees (Goodkind, Jessica R.; Hang, P.; Yang, 2004; Goodkind, 2005; Goodkind et al., 2014, 2020).

Immigrant Well-being Project (IWP): Partnership to Adapt the Model

From the outset of the IWP study to adapt RWP for Latinx/@ immigrants, the intent was to work with existing Latinx/@ organizations to build on their knowledge and experience and avoid duplicating efforts, with the hope that the collaboration would be sustainable and have synergistic, multiplicative effects. The IWP was established as a partnership comprised of one university and four organizations working in different domains to meet the needs of Latinx/@ immigrants: Centro Sávila provides high quality mental and behavioral health services regardless of insurance or ability to pay; Encuentro provides adult educational classes for English language learners, and for certification and professionalization in areas in which immigrants already have skills, including small business development; New Mexico Dream Team (NMDT) empowers youth and mobilizes policy changes for immigrant families; and the New Mexico Immigrant Law Center (NMILC), provides free and low cost legal services for immigrants within a framework of community empowerment.

This article describes the initial phase of research to adapt and implement the IWP, which consisted of conducting an in-depth study of the mental health needs, stressors, current political/economic/social context, and local solutions as experienced by 24 Latinx/@ recent and non-recent immigrants and their families to inform the adaptation of the RWP model for IWP. IWP is ongoing and is testing feasibility and preliminary outcomes of the intervention to improve quality of life and well-being of Latinx/@ immigrants. Our aim in this article is to show how the CBPR processes developed by the community-university research team to collect, analyze and, disseminate the data from the initial ethnographic interviews were developed collaboratively, describe their innovative features, and highlight how they advance participatory methods and our understanding of the centrality of language in research processes. To carry out the collaborative processes described in this article, the IWP research team of university and community organization members met twice per month to engage in all research activities together. In addition, we established a broader community-university advisory council that included leaders of the community organizations and community members who were clients of community partner organizations.

Data Collection

Twenty-four ethnographic interviews were analyzed using the participatory processes described in this article. For description of the ethnographic interview participants, data collection procedures for the interviews, and associated findings, see Vasquez Guzman et al. 2020b. Here, we describe the participatory research meetings and retreats we engaged in as a team to analyze the interview data. To track our participatory research process, we analyzed meeting and retreat notes. Notes were taken by one team member for each research meeting and multiple team members at each retreat. Notes were taken in Spanish and English, depending on the language of speakers. During small group sessions in the retreats, a note taker was designated, or notes were actively taken as part of the small group process, and were relayed to the larger group in summary form. All notes for each retreat were then integrated and distributed to the larger group for feedback.

Data Analysis

The focus of this paper is on the six stage participatory data analytic process that we developed and implemented: 1) Establishing the Community-University Research Team, 2) Planning the Interview Process/Data Collection, 3) Developing the Data Analysis Methodology, 4) Interpreting Findings to Adapt the Intervention, 5) Integrating the Results of the Participatory Process into the Analysis, and 6) Data Analysis for Dissemination.

Stage 1: Establishing the Community-University Research Team

The study began in September 2017. Together, we established a schedule of biweekly research meetings and a community-university research team composed of one or two members from each community organization, as well as academic partners. Faculty members included the PI (a white community psychologist with over 20 years of experience partnering with refugee, immigrant, and indigenous communities to create and study community-based mental health interventions), and four co-investigators (a bilingual Latina epidemiologist whose research concerns maternal-child health; a Diné (Navajo) sociologist whose research examines health equity; a white Spanish-speaking cultural anthropologist whose work concerns immigrant and refugee belonging and well-being; and a native Spanish-speaking Mexican-American immigrant sociologist who studies health policy issues that affect Latinx/@ people). The team also included two doctoral students from sociology and clinical psychology, an undergraduate student in business, all of whom are Mexican-American, as well as a Latinx/@ research coordinator. Community organization research team members included a licensed mental health counselor and community health worker from Centro Sávila, a community health worker from Encuentro, a youth leader/community organizer from the NMDT, and an attorney from NMILC. All community research team members are Spanish speakers, or bilingual in Spanish and English, and are either immigrants themselves or from immigrant families. Community partner organizations had subcontracts with the university ($15,000 per organization during Year 1 of the study when these analyses took place and $5,000-$15,000 per organization in each subsequent year, depending on the level of involvement), and thus the organizations and staff members were paid for their participation in the research processes.

Finding a regular meeting time for all participants was challenging, and 1½ hours twice a month felt rushed for such a large group to listen, build trust and to develop plans together. Thus, we decided to try a half-day (5 hour) retreat to provide more time for more in-depth learning and reflection. The retreats included not only the community-university research team, but also leaders of the community organizations and community members who used services or participated in the four organizations, but were not paid staff. Community member participants who were not organization staff received $100 honoraria for their participation in each retreat. Over the course of the initial research phase described in this article, the team held six retreats. The retreats were held in conference rooms or other spaces made available by community partner organizations. Food was provided, breakfast at the beginning and a working lunch at the end.

The first retreat was held in December 2017. At that retreat, members of the university team shared the RWP model and outcomes. Each community organization shared its mission and activities. We began to talk about how to adapt the RWP model for immigrants and our partnership structure. Members appreciated the opportunity to understand the RWP model more thoroughly. Learning about the work of the community organizations was key to begin to understand the strengths each brought to the partnership in terms of knowledge of the challenges and strengths that immigrants face in the community, but also the necessity of flexibility, given the swiftly changing policy and social context around immigration. This approach demonstrated the value the partnership placed on mutual learning from the onset. Important themes emerged from this beginning, including: incorporating a focus on youth, educating IWP participants about community resources regardless of legal status, and addressing fear and/or stigma that surrounds seeking services. The emerging youth focus led to discussions about IWP becoming a vehicle for youth leadership development, with one partner noting that we needed to “help community members reach a level where they feel safe enough to act as agents for change and advocacy.”

The majority of research team members were bilingual. A few were dominant Spanish speakers and there were two members of the team (PI and a co-investigator) who did not speak Spanish. Many grew up speaking Spanish at home, but were mostly educated in English and so often found it easier to speak English in work or academic settings. Others learned Spanish as adults, but were functionally bilingual. During our first meetings, we primarily spoke English, and translated into Spanish when necessary. In an effort to be more inclusive and to not privilege English, we began experimenting with language processes when we met. We attempted a model where people could speak whatever language they were comfortable in and interpretation would be provided for those who needed it. Thus, from the beginning of our research processes, we recognized the importance of language and the power it dictates. We also recognized the utility of intensive 5-hour research retreats every three months (in addition to our twice monthly research team meetings) for both making progress on our research goals and building trust and relationships within the partnership.

Stage 2: Planning the Interview Process/Data Collection

Our team actively created opportunities for mutual learning to arrive at a data collection plan that included and built upon academic and community partners’ knowledge. The retreat structure provided time and space to review and co-create data collection tools that better captured the complex lived experiences of Latinx/@ immigrants. During the second retreat in February 2018, we continued our focus on process and research goals. We worked with CBPR expert Dr. Nina Wallerstein and her team to do collective visioning of our partnership using their “river of life” tool (Sanchez-Youngman & Wallerstein, 2018). This provided the opportunity for partners to share their individual and organizational histories and chart a collective plan for the future together, which resulted in deepening of our relationships. We also discussed the process of conducting ethnographic interviews, including how we would recruit participants, consent processes and scripts/forms, and reviewing and selecting interview questions. We agreed upon obtaining verbal, not written, consent from participants, and revised the consent form to more clearly convey the fact that we had a Certificate of Confidentiality that protected participants’ identities from subpoena from federal agencies. Both of these measures were important given the context of criminalization of immigrants without documentation.

This iterative work began in the second retreat but was completed over the following month at our team meetings and via email to ensure that consenting and interview processes reflected the input of all community-university research team members. Documents were reviewed in English and Spanish to further ensure consensus. This was imperative because often during the translation of various instruments and surveys, meaning can be lost and/or misunderstood. Establishing that the team was flexible and open to community organization representatives’ input was essential in developing the interview guide. The team began with questions from previous studies we had submitted in the funding and institutional review board processes. However, these questions were substantively changed through our discussions during the retreat. For example, beginning from a position that we did not want to re-traumatize people by asking about migration experiences, community representative team members strongly advocated including these questions as essential to people’s experiences, however, also ensuring that participants did not feel compelled to share responses to any question. Having community input was invaluable as the importance of lived experiences and making sure that questions would be relatable to participants was emphasized. Furthermore, when possible, community partners participated in co-conducting the interviews, which was feasible because the interviews were conducted in Spanish.

Stage 3: Developing the Data Analysis Methodology

The third retreat was held in June 2018. We decided to translate a portion of the data (four interviews) into English, so that English-dominant speakers could participate in the data analysis. The goal was not to replicate academic coding and interpretation processes; rather to develop a process that was rigorous and inclusive and allowed space and ample time for the process to unfold. A colleague recommended Hallett, Held, Knows His Gun McCormick et al.’s (2017) article, “What Touched Your Heart? Collaborative Story Analysis Emerging from an Apsáalooke Cultural Context” that described the participatory processes developed in their partnership between the Apsáalooke Nation and Montana State University. Their methods allowed indigenous epistemologies to emerge in the process. This seemed important to try to replicate in this context. We gave a transcript to each research team member in their preferred language, explaining that “What we are trying to do is to pay close attention to what the interview participants told us and figure out how to incorporate what we are learning into the intervention model that we will develop.” Additionally, we asked research team members to consider three questions as they read: 1) What stood out to you in the person’s interview?; 2) What touched your heart?; and 3) If you were to develop a program/intervention for this person to help improve their health and well-being and assist them in their settlement process, what would that look like?

Asking research team members to reflect on “what touched their heart” as they read was part of an effort to distance ourselves from biomedical or Western understandings of what is important, and to make sure participants understood that affective responses were encouraged as well. This part of the process underscored the ways that language is feeling and how moving beyond a traditional analytical framework allowed for lived experiences and voices that are often marginalized to come to the forefront. Moreover, because immigration is an emotion-laded topic, the process worked to humanize participants, highlight the way current policies and practices exclude them, and to illuminate discrepancies between lived experience and the national discourse on immigration.

Building on Hallett and colleagues’ method, we added the important component of keeping the transcripts in participants’ native language and incorporating the feelings evoked from interviews into our analysis. The result of this process was an in-depth analysis that revealed layers of meaning made possible by the genuine participation of all members and the role of researchers’ affective responses in data analysis.

In addition, the retreat produced a list of themes that served as the basis for a code book. However, the way these emerged was through a broad discussion of different sub-groups and intersectional identities within the Latinx/@ immigrant community whose needs and perspectives should be addressed. A university team member noted that the young man whose transcript she read stated, “‘I don’t get sick.” He said he didn’t want to see a counselor and went to the gym instead. We need to think beyond counseling as a way to reach these kids.” A NMDT member said we need to intentionally include the experiences of immigrant women, stating “the violence that these women experience can be extreme. We don’t talk about these things and every woman has a story. The stories of LGBTQ communities are also important.” Another participant said that this was an “opportunity to understand the experiences of younger generations of immigrants.”

This led to a discussion of the importance of intergenerational learning, and one university researcher noted that RWP Learning Circles were a “great place for families to have a space where they feel respected by their kids and parents” and that this could be replicated in the adapted intervention. A research team member from NMILC suggested that interviewees “should have a space to share with one another and meet one another. We could share the themes that came out today. That would be really powerful.” As these themes and illustrative quotes demonstrate, we came away from the retreat with a thematic structure that we could use to begin to apply to the data, a set of topics to address in the intervention, and a plan for continuing our analysis by including interviewees at our next retreat.

During these longer retreats, participants spoke in the language they were most comfortable in, sometimes switching between Spanish and English, and interpretation was ongoing to keep the group on the same page. This process allowed us to learn from each other and ensure that all voices were heard, thereby increasing the efficacy and power of the team.

Stage 4: Interpreting Findings to Adapt the Intervention

The fourth retreat occurred July 2018, only one month later, because the start of the intervention with immigrant families was imminent. The goals of this retreat were to continue the data analysis process and to plan for intervention implementation in November. Several aspects of this retreat deepened our collaboration: we invited several interview participants to attend the retreat and we decided it was important to listen to interviews in addition to reading them to better understand the emotional valences and context of interviewees. We asked two interviewees for permission to listen to their audio file in a group setting prior to the meeting.

While listening to portions of two selected interviews, group participants were invited to discuss primary themes, many of which became essential to our coding tree, including “no saber/not knowing, or uncertainty,” “health care access and resource seeking,” “collective wellbeing,” and “using education to confront obstacles.” The research team’s goal in listening to the audio recordings was to reinforce the importance of paying attention to affect and feeling, which can be lost when reading transcripts. After listening, many participants did remark on the emotional impact of hearing participants’ voices. However, even more notable was when a research team member offered interpretation of participant experiences, and an interview participant “spoke back” to that interpretation, immediately offering a more complex, nuanced interpretation. Because the group had been practicing participatory processes for many months, the immediate consideration of this perspective which thereby served to incorporate it into the analysis, showed that the group took participant perspectives seriously. Further, these interview participants continued to participate in future retreats becoming part of the larger team contributing to analysis. Our focus on privileging the language of interviewees enabled us to engage in genuine collaboration throughout the data analysis and intervention development processes.

Stage 5: Integrating the Results of the Participatory Process into the Analysis

Our fifth retreat, in September 2018, was dedicated to finalizing the intervention model. As focus shifted to implementing the intervention, the research team completed coding all transcripts. Departing from a process used by the first author to develop a coding tree with academic research team members, the team began by using the thematic categories developed during the third and fourth retreats. Themes were entered into NVivo as the initial codes and marked to indicate that they were collaboratively produced. Four university members of the coding team proceeded to code transcripts and finalized the codebook for several weeks, then met weekly to re-name, merge, delete and define nodes. The codes were largely descriptive so as to encourage intercoder agreement and leave interpretation and analysis for further stages. The codes were bilingual, beginning with Spanish, followed by English translations, for example, “Salud, Health.”

In order to include community team members in the continuing data analysis, we invited them to participate in memo writing. Memo writing was introduced as a way of further interrogating, exploring and interpreting data gathered in descriptive codes (Miles et al., 2013). Because focused coding and memo writing are features of constructivist grounded theory methodologies that emphasize co-construction of interpretation and analysis (Charmaz, 2014), they are amenable to collaborative processes. The Spanish-language data and bilingual coding enabled community research team members who were Spanish-dominant to fully participate in the process. The third author trained four community research team members in memo writing techniques and led the group in determining priority themes to explore through memo writing. Memos were written on prominent themes and themes the group decided were key to the analysis. Memos explored the extent of a theme and the range of meanings, allowing further definition and exploration of each theme and how it connected to other themes. Analytical processes become more transparent when they are systematic and all team members participate in them using concepts derived from everyday language. Next, all nine community university research team members who helped with memo writing met for a data analysis mini-retreat in May 2019.

During the May 2019 mini-retreat, our goal was to develop a framework for a results paper (Vasquez Guzman et al., 2020). Each team member read three or more memos of 30 completed memos that summarized and interpreted all of the text coded at themes including “stress,” “well-being,” “health,” and “family separation.” We spent 45 minutes reading, jotting notes, new ideas or directions to follow, as well as highlighting illustrative quotes. We spent most of the retreat sharing our insights gleaned from the memos we read, while starting to think of “big picture” themes that would provide an analytical framework for the results. What emerged was the importance of understanding each participant’s life trajectory as a story of inclusion, exclusion, or both, with important subcategories including health, legal and policy, language, finances, and aspects of identity (e.g., sex, gender, age, status), and a particular focus on immigrant agency and structure as it propelled each person’s story with a set of constrained choices at multiple levels. After the retreat, each community-university research team member read through an assigned number of interview transcripts and created a summary of each participant’s life trajectory and immigration experience.

During this process, community members related to the stories they read. Hearing the stories and connecting themes as they emerged across the stories was transformative for our research team members. Because the team included Spanish-speaking immigrants whose own lives had been affected by the current social-political climate, in documenting participants’ trajectories and how different community organizations made a difference within their lives, the team’s awareness and understanding of how transformation occurs was reinforced.

Stage 6: Data Analysis for Dissemination

Once the participant life trajectories were completed, the team met and agreed that our findings could best be interpreted and understood in terms of different trajectories using a theoretical framework of mobilities. A mobilities framework is often applied to contexts of transnational migration, but less commonly used in post-migration contexts. Thus, the third author led the team in conceptualizing movement or mobility between statuses of inclusion and exclusion in the United States. These theoretical and analytical models served to clearly highlight what the team had seen in the data: that individual agency within particular social contexts and policies shaped immigrant inclusion and exclusion. The team collaborated on an academic paper based on these results (Vasquez Guzman et al., 2020) and a research brief in both Spanish and English to further disseminate these results. In addition, the team has given bilingual presentations of these results at several conferences and community events.

Table 1 summarizes our participatory methods, highlighting how they illustrate methodological praxis, their innovative aspects, and the outcomes with respect to the relationship between language and these transformative processes. Each stage made the next one possible. Our results thus show how our processes that centered participants’ language unfolded to build trust, co-create stronger data instruments and collection processes, and improve the rigor and validity of our data analysis, results, and dissemination.

Table 1.

Participatory Bilingual Research Process Stages

Research Stage	Objective	Innovation / Praxis	Outcome
Stage 1: Establishing the Research Team	To establish a research team inclusive of community and university members	Hold research team meetings in Spanish; put research-language into understandable terms for community members to fully be a part of the team	Language is Power: Disrupted power dynamics - monolingual English research team members had to ask for interpretation - and contributed to decolonizing research frameworks
Stage 2: Planning the Interview Process/Data Collection	To plan data collection, including creating interview guides and conducting interviews by prioritizing community partners	Construct the interview guide in Spanish to ensure that meaningful aspects of immigrants’ lives were asked about in language they would use before translating into English for IRB review	Language is Meaning: Interview guide was more meaningful and better able to reveal the complex lived experiences of participants
Stage 3: Developing the Data Analysis Methodology	To develop systematic, rigorous data analysis that bridged academic and participant ways of knowing, including a focus on affect	Use university and community research team members’ emotional responses to the transcripts to guide analysis of interview data - emphasis on heart rather than head AND integration of both intertwined aspects of meaning-making	Language is Feeling: Recognition of team members’ own experiences of trauma and stress and creation of space for healing and support built trust and stronger research-community partnership. Disrupted false dichotomy between reason and emotion
Stage 4: Interpreting Findings to Adapt the Intervention	To incorporate community interpretations into analysis.	Listen to audio files in presence of interview participants who spoke back and extended interpretation to deepen collaboration and analysis	Language is Collaborative: Participants’ reactions were immediately incorporated into analyses and results. Extended processes for incorporating all team members, including research participants, into data analysis
Stage 5: Integrating the Results of the Participatory Process into the Analysis	To include community organization research team members in analysis processes, including memo writing and model-building	Conduct data collection and coding in Spanish. Write analytical memos with community and university team members collaboratively	Language is Analytic: Collaborative memo-writing led to creation of innovative, community-grounded theoretical and analytic model: mobilities and trajectories of inclusion and exclusion
Stage 6: Data Analysis for Dissemination	To disseminate results to academic and non-academic audiences	Co-created bilingual presentations in multiple forms and delivered them to multiple audiences	Language is Transformative: The innovative, participatory, bilingual research processes transformed university and community research team members, their relationships with each other, the research team as a whole, and the research enterprise

Results & Discussion

Although CBPR researchers have generally accepted that community involvement improves the rigor and outcomes of research, there remains a need to understand how community partners can be meaningfully included in data analyses and how such inclusion strengthens the process of inquiry (Blumenthal, 2011; Committee, 2011). We developed a series of six stages of concrete practices through which participatory data analysis, interpretation, and dissemination can occur. Descriptions of these stages and their outcomes for our research collaboration demonstrate how a focus on praxis and language can have important implications for participatory research in bilingual or hybrid language contexts. Through collaborative interpretation of these outcomes, we gained insights into the multiple roles of language in research and how our development and implementation of innovative participatory bilingual processes ultimately led to transformation among all members of the community-research team and the research process itself.

Stage 1: Language is Power

Spanish is the dominant language of the participants in this study. However, the everyday lives of immigrants living in the United States are characterized by hybrid, bilingual language practices (Field, 2011). Although we designed the study with the intention of collecting data in Spanish, the participatory principles that guided the research prompted us to explore language practices in the research team context as well. With respect to meetings—a key element of research team praxis—we experimented. We ultimately decided on the principle of participants speaking the language they were most comfortable with, often resulting in Spanish-dominant meetings, since the majority of the team was bilingual, and the inclusion of dominant Spanish speakers resulted in Spanish-dominant meetings and interactions. In practice, this meant that two senior members of the team had to request interpretation. English, the dominant language in Western-based research paradigms, was de-emphasized. This process, however, was not a simple reversal from English to Spanish, because even bilingual members of the university team who were native Spanish speakers had learned about research primarily in English, so these inclusive practices necessitated not just the learning of Spanish-language vocabulary for research terminology, but also figuring out ways of describing how these were put into practice in everyday language. This disrupted the traditional power dynamics of the research team and expanded opportunities to decolonize research practices.

Reflexivity about researcher positionality and power in research that has transformative aims is an important part of this process (Giampapa & Lamoureux, 2011). Although there has been a call for the importance of these types of language practices, research that specifically describes examples of these methods and how they work in practice is limited. Furthermore, our methods highlighted how language is power because they shed light on how language is a form of constructing one’s social reality (Carter & Sealey, 2000). Academic frameworks and jargon have long dominated research production and what constitutes knowledge, but when multiple languages are included and methods are co-created, they more closely mirror the multiple voices and perspectives of communities. Our active prioritization of the native language spoken by members of our research team resulted in deconstruction of English-dominant research frameworks and a greater voice for community organization staff and community members.

Language is a critical driver of immigrant economic and social integration into the United States (Portes & Rumbaut, 2014). Similarly, we found that language is also an important driver of immigrant contribution and engagement in the research enterprise. In the research context, language can either facilitate or hinder the degree to which immigrants not only take part in research projects, but also are able to actively participate in knowledge production through data analysis, interpretation, and dissemination. Research has shown that employers who include immigrant workers in workplace decision-making often see higher levels of return, and inclusivity often translates to better products and outcomes (Hunt et al., 2019). While the corporate world has done a better job at tailoring their products for diverse audiences, the research enterprise has remained English-dominant and centered in Western knowledge frameworks and approaches.

Stage 2: Language is Meaning

Language orders experience in the world. From the beginning of the discipline, ethnographers have recognized the importance of participant observation and conducting research in participants’ native languages (Boas, 1940; Malinowski, 1922). Building on these long-established practices, we wanted to apply them in the complex, hybrid-language research space our participants and research team occupied. The half-day research retreats afforded time to engage in collective visioning and allowed for rapport- and trust-building. Beginning with discussion of consent processes and working together to find compromises for accommodating NIH, university IRB requirements and participants’ concerns, we demonstrated openness and commitment to upholding the views of each team member. Thus, when creating the interview guide, there was thoughtful, engaged response and efforts to make certain the guide captured everyday experience, the positive and negative aspects of transnational lives, experiences of many kinds of violence (e.g. intimate partner, structural, state), and hopes and aspirations for the future.

Health researchers have advocated for employing the culture concept as a critical feature of research methods, which includes going beyond mere translations of concepts and constructs (Kao et al., 2004). Falicov (2009) has written about the importance of bilingual/bicultural adaptations for Latinx/@ immigrants. Conducting research conversations in Spanish ensured that explanations of the purpose of the interviews and consent processes, as well as the actual interview questions, were grounded in participants’ experiences and perspectives. We were not just translating questions developed by researchers from English to Spanish, but developing them from the ground up based on community member/researcher perspectives. This also reflects the importance of prioritizing Spanish during the co-development of the interview guide, enabling the team to arrive at an interview guide that was more meaningful, better capturing the complex lived experiences of participants. These processes highlighted that language is meaning, underscoring the ways in which meaning emerges from language and symbols (Goffman, 1959, 1967; Schwalbe, 1983), and because language also emerges from practice, it was co-created in the research space. Our team’s construction of meaning directly evolved from the fact that our social reality was grounded in Spanish but also the cultural customs and ideology of Latinx/@ immigrants’ cultures.

Stage 3: Language is Feeling

Researchers have been more attuned to emotion, feeling, and affect in recent years, a focus that Clough identified as the “affective turn” (2006). This has shaped ethnography (Behar, 1996), migration research (Boehm & Swank, 2011), and participatory approaches that lead to social transformation (Langhout, 2015). It is especially important in team science that is inclusive of members who may have experienced the same traumas explored by the research to be attentive to addressing feelings in ways that promote healing. Through collecting data on violent and traumatic experiences related to pre- and post-migratory contexts, and on the migration journey itself, we knew we must also attend to interviewer responses to the interview process. Recognition of team members’ own experiences of trauma and the creation of space for healing built trust and a stronger partnership between academic and community members.

Moreover, we set the stage for attending to affect in the interpretive and analytic by the use of Hallett et al.’s (2011) process to elicit community-ways of knowing. This approach was compelling because of its explicit attention to emotion and heart. While other existing interpretive frameworks are firmly rooted in Western epistemologies (Smith, 2012), Hallett and colleagues’ work offered a framework to elicit feeling rather than falling back on the false dichotomy between knowledge and emotion. The question “What touched your heart?” created a space for team members to share what moved them, and thus kept our goal of recognizing and challenging inequities at the forefront of our efforts.

Feeling has long been acknowledged as an aspect of language development and expression (Bloom & Beckwith, 1989; Goldberg, 1971), however, attention to feeling in research practice was, until recently, less developed. Asserting that language is feeling is only the first step in incorporation of feeling into research practice. Being attuned to both mind and passions, the ability to act and be affected, are now understood to be critical to epistemology. Participatory processes therefore must incorporate mechanisms to elicit feeling and include it, not only in interpretation of data, but also in the analytical models.

Stage 4: Language is Collaborative

Eliciting participant responses to research has previously been incorporated into ethnographic methods (Brettell, 1993). Across social science, this process has become known as “member checking,” and is often touted as a way to enhance validity and trustworthiness of qualitative research (Creswell & Miller, 2000). Therefore, although many of the processes described here are well established in ethnographic, qualitative, and participatory research, and although they may enhance or suggest the possibility of collaborative practices, the primary goal of “member checking” has traditionally been enhancing trustworthiness and ensuring audiences of the validity of the data. Adhering to Spanish as our primary language of analysis enhanced the collaborative nature of the relationship between community and university research team partners because it enabled inclusion of participants on the research team at multiple levels and enabled seeking input more collaboratively. Additionally, the combination of using emotion as an epistemological feature and the collaborative interpretation during retreats allowed space for research participants to “speak back” and was a stark reminder of the personal nature of the research process. It served as further impetus to keep close to the data and our overall goal of creating interventions that addressed mental health inequities on multiple levels. It also facilitated the ability to incorporate participant responses and interpretations directly into the analysis. Further, the multiple and varied interpretations by team members and participants of the same scenario emphasized the importance of attending to intersectionality and the multiple axes of difference (e.g., age, generation, sexual orientation) in both the research team and the intervention. In sum, we found that language is collaborative because genuinely bilingual research methods facilitated a dialogue and dialectic at multiple levels of the research team and research design.

Stage 5: Language is Analytic

Memo writing is a cornerstone of qualitative analytic processes (Charmaz, 2014; Miles et al., 2013). Keeping the data in Spanish was critical to enabling those primarily literate in Spanish to participate in these processes in a team setting. Having community organization representatives participate in analytical processes served to deepen the research processes through describing models, processes, and relationships in the language participants used, and using concepts that explain social dynamics that made sense to Latinx/@ immigrants in everyday life. This dynamic can be described as productive in that it produced analytical constructs, but it also required checking through translation to make sure constructs first conceptualized in English made sense when translated. Thus, “mobilities” as a theoretical framework was both produced and checked in a collaborative fashion. The analytical model that was produced—trajectories of inclusion and/or exclusion—came directly out of our discussions of the data and the case studies produced by each team member in Spanish. In this way, language is analytic in that attention to language in the co-creation of analysis worked toward creating the alignment between researchers’ and participants’ social worlds described by Eisenhart (2011). Moreover, we found that this process resulted in not only more robust and valid data as suggested by Fox and colleagues (2019), but also analytical models that both advance research and interventions to address health disparities.

Stage 6: Language is Transformative

The IWP team, from its inception, knew we had to reach beyond academia for social transformation and justice. Having community partner organizations actively involved in every stage ensured that collaborative processes would inform not only design and data collection, but also interpretation, analysis and dissemination. Our collaborative and inclusive approach led to transformation for participants and for community university research team members. The stages of research are iterative, but they become much more complex and meaningful when involving community voices and perspectives. The lived experiences, the stories, and the unique social locations of Latinx/@ immigrants were included from multiple perspectives leading to transformational changes personally and intellectually. The four community partner organizations’ social capital, leadership, and data analytical skills all improved. These are the processes which foster sustainable change, not just empowering or improving the health of research participants, but by building capacity of community and university researchers who learned research vocabulary in Spanish and developed collaborative skills and innovative methodological approaches.

Furthermore, the collaborative bilingual approach extended to dissemination of research findings at conferences. The presentations were bilingual to ensure to provide that platform to enable community partners to share the team’s progress and our results. Bilingual presentations also served to illuminate how language use intersects with power dynamics, as monolingual audience members experienced disruptions when listening to a presentation in a language they do not understand and waiting for translation. Attending conferences was also transformational all community-university research team members by providing opportunities for further teambuilding, networking, and professional development.

The intentional focus on multiple aspects of language (its power, different meanings, feelings evoked, and as a tool of collaboration and analysis) has highlighted that language is transformative in capacity and relationship-building among all community-university research team members, as well as in the multi-level intervention we co-developed and co-implemented.

Limitations & Implications of Approach

This work has important implications for the advancement of CBPR methodology. CBPR research has shown how including participants in all phases of research strengthens rigor (Vaughn et al., 2017). However, most CBPR studies document participation in recruitment and data collection phases. We add to the growing body of literature that meaningfully includes participants in data analysis. In addition, this study offers detailed description of the processes of inclusion that incorporated participants’ language and ways of knowing, and thus transformed the research process itself. In addition, our collaborative processes that focused on praxis and language show how power imbalances inherent to research can be brought to the fore and addressed during the course of research praxis.

Moreover, incorporating participants as co-researchers built capacity with respect to skills and knowledge of all community-university research team members. Language skills related to technical vocabulary and ability to explain research concepts improved. In addition, developing innovative analytical processes advanced science by extending traditional methods to incorporate local epistemologies that reflect complex social realities. Our focus on dissemination resulted in gains in practical skills: developing bilingual presentations and public speaking for a variety of audiences.

Although language was a powerful tool in our research, we do not suggest that we fully transformed unequal power relations among research team members nor that our interactions were without challenges. The bilingual, but Spanish-dominant language structure that we utilized was not ideal for all partners all of the time, but everyone demonstrated flexibility and patience. For example, community research team members sometimes felt overwhelmed by vague or complex e-mail messages, especially if they were sent in English. Additionally, we found that truly inclusive research practices often double or triple the typical time it might take for tasks such as coding and memo-writing, which is due to interpretation/translation. The term “second shift” (Hochschild & Machung, 2012) might apply for some bilingual university team members who engaged in this cultural brokerage work and bore the brunt of these responsibilities.

Next steps include further expansion of these methods in the implementation and evaluation phases of the intervention study. Detailed documentation of these processes and how they might be adapted to local conditions and organizational contexts is important. Multi-level system change efforts need broad dissemination plans that include community level and policy-maker audiences. Thus, IWP is also working on developing bilingual policy briefs using similar participatory collaborative processes and will continue to actively explore dissemination of our processes and results using a wide variety of channels that extend beyond academic publishing.

Conclusion

In the paper, we described the meaningful ways community partners were involved in data analysis and dissemination in a CBPR study. While a growing body of research utilizes CBPR approaches with immigrants (Vaughn et al. 2017), this study provides a roadmap for large-scale multifaceted research projects to focus on research processes as well as social change outcomes to improve the mental health of Latinx/@ immigrants. We found that attention to the specific methods of data analysis led to transformation for both community and university team members. Language is often employed to shape ideas about immigration and immigrants as constituting threats to the United States. Although many have been vocal to counter these hateful and divisive narratives, mental health researchers play an important role in understanding the impact of hostile contexts on immigrant well-being and on changing the discourse. Here, we reclaim language and use it as a means to engage, empower, and transform data analysis processes in research and intervention. If we embrace the notion that health equity is an essential goal in mental health intervention research, then we must incorporate processes that attempt to address inequities in the research itself. Through deliberate construction and implementation of an innovative participatory bilingual data analysis effort, our CPBR team found that language and methodological praxis can be used as tools for achieving genuine inclusion and social justice in research with Latinx/@ immigrant communities.

Public Significance Statement:

Employing participatory approaches on a diverse research team comprised of community and university members leads to rigorous research that can better achieve health equity for Latinx/@ immigrants. This research demonstrates how our innovative collaborative approach that involves all team members in data collection, interpretation, analysis and dissemination has a high potential for contributing to interventions that will increase health equity. Furthermore, by prioritizing Spanish and focusing on language we actively deconstructed traditional frameworks for research thereby empowering both research and community members.