Hmong Older Adults’ Perceptions of Insider and Outsider Researchers: Does it Matter for Research Participation?

Abstract

Background

Recruiting racial/ethnic minorities into health research is challenging. While researchers affiliated with members of the study population (seen as insiders) may increase research participation of racial/ethnic minorities, little is known about who participants see as insiders and how they respond to insider versus outsider researchers.

Objectives

To examine perceptions of Hmong older adults toward insider versus outsider researchers and the impact of these perceptions on their willingness to participate in research.

Methods

Participants in a study evaluating use of a culturally and linguistically adapted audio computer-assisted self-interviewing system with helper assistance (ACASI-H) provided information about what they thought would encourage others in their community to participate in research. ACASI-H was used for collection of health data with 30 Hmong older adults. Participants rated likelihood of participation if the researcher was Hmong and answered open-ended questions about participation when the researcher was not Hmong. Conventional content analysis was used to analyze open-ended questions.

Results

Sixteen (53%) participants reported they would be “likely” to participate in the research if the researcher was Hmong. Fourteen (47%) participants reported they would participate in research if the researcher was not Hmong. In addition to ethnic affiliation, trust in the researcher could shift the perception of the researcher toward insider status, thereby increasing willingness of Hmong adults to participate in research. Trust in the researcher, and movement toward insider status, could be increased by calling out a connection between the researcher and participant or creating reciprocity.

Discussion

Findings suggest that increasing research participation of Hmong (and possibly other) participants, particularly when researchers does not share ethnic membership, can be achieved by building trust. This study also suggests a more nuanced perception of insider status, as a continuum rather than a dichotomy, may be a more accurate reflection of the relationship between participants and researchers.

Background

Recruiting and retaining racial/ethnic minority populations into research is challenging (George, Duran, & Norris, 2014; Yancey, Ortega, & Kumanyika, 2006). Consequently, racial/ethnic minorities remain underrepresented in research, thereby limiting researchers’, clinicians’, and policy makers’ capacity to tailor care and develop policies to improve health outcomes for them. Barriers to recruiting and retaining racial/ethnic minorities include: the historical context of research in minority communities (e.g., African-Americans who took part in the Tuskagee experiment); participants’ mistrust of researchers; poor understanding about or interest in research activities; fear of deportation due to immigration status; and researchers’ lack of attention and sensitivity to language and cultural differences (George et al., 2014).

Researchers have developed strategies to address these challenges, including community engagement activities (e.g., community outreach, community advisory board), working with community professionals and/or community members who are part of the study population to refer participants, and targeted marketing (e.g., using flyers and social media; (Chadiha et al., 2011; Horowitz, Brenner, Lachapelle, Amara, & Arniella, 2009; Langford, Resnicow, & Beasley, 2015; Mendez-Luck et al., 2011). For example, one study used community outreach strategies including board-sponsored (e.g., local residents, activists, and leaders) special recruitment events such as live music, dancing, and giveaways, as well as public events such as health fairs, resulting in enrollment of 68% (377/555) African and Latino participants (Horowitz et al., 2009). While some strategies have been effective, researchers have not collected systematic data on recruitment barriers or compared the success of different strategies to determine which are most effective at increasing recruitment, thus limiting generalizability.

One theory for approaching this problem that has received considerable attention is increasing the number of academics studying their own (racial/ethnic) communities (Espenshade, 2001; Kusow, 2003). This practice of matching ethnic backgrounds of researchers and participants has been increasingly used across disciplines and is generally described as the insider versus (vs) outsider debate.

Insider vs Outsider

Insider vs outsider status has been studied since 1972 by sociologists including Merton (1972) and anthropologists including Headland and colleagues (Headland, Pike, & Harris, 1990). Insiders have been defined as “members of a specified group… or occupant of a specific social status [within that group]” (Merton, 1972, p.11). In consideration of Merton’s (1972) work, researchers have examined the epistemological and methodological significance of insider vs outsider status for researchers (e.g. Breen, 2007; Greene, 2014). Some have positioned insider vs outsider identities as opposites, both with advantages and disadvantages. For instance, benefits of being an insider include having a greater understanding of the study group culture, having an established relationship with the community, being able to establish rapport with study participants, being sensitive to the community’s culture and needs, and dealing effectively with ethical concerns (Blythe, Wilkes, Jackson, & Halcomb, 2013; Burns, Fenwick, Schmied, & Sheehan, 2012; Greene, 2014). There is assumed credibility, quicker access, trust-building, shared sociocultural experience, and an ability to communicate with non-English speaking participants in their native language (Blythe et al., 2013; Burns et al., 2012; Greene, 2014). Some argue, however, that each benefit is linked to a disadvantage such as threats to validity of the study findings. For example, insider researchers’ familiarity with participants can lead to a loss of “objectivity” and assumptions about the data based on their prior knowledge rather than the actual results (Blythe et al., 2013; DeLyser, 2001). This extends to concerns that insider researchers may have more difficulty gathering data because the study participants might assume the researcher already knows the answer (DeLyser, 2001). Some have suggested that sharing an ethnic background might not confer insider status, questioning of how insider vs outsider status is determined since there are many ways to be different (Breen, 2007; Ochieng, 2010).

Outsider researchers have indicated there are benefits in being naïve to the study population, such as observing participants’ actions without judging them and providing an objective account (Adler & Adler, 1994; Innes, 2009; Knobf et al., 2007). However, disadvantages for outsiders include longer time to gain trust, limited cultural insights, and lack of acceptance by the community (Hodkinson, 2005; Innes, 2009).

This view of insider vs outsider identities as opposites does not take into consideration how identities are interconnected and that they are not static (Hall, 1990; Wray & Bartholomew, 2010). Furthermore, the literature on insider vs outsider researchers is comprised mainly of researchers reflecting on their own experiences, their perceptions of their positions as insider or outsider, and how they believe their position affected recruitment and data collection with participants from their own cultural background. Other important questions remain unaddressed such as; how participants perceive researchers who share their ethnic (or other) status and how their perceptions of the researchers impact their decision to participate in research. Therefore, the purpose of this paper was to explore perceptions about insider vs outsider researcher status and the impact of these perceptions among Hmong people’s willingness to participate in research.

The Hmong

There are 260, 076 Hmong, a Southeast Asian minority group, living in the US. They immigrated to the US as refugees after the Vietnam War because of their support to the Central Intelligence Agency’s efforts in Laos during the Vietnam War and experience of persecution by the Pathet Lao communists after the war ended (Hamilton-Merritt, 1993). The Hmong are most concentrated in California, Minnesota, and Wisconsin (Pfeifer, Sullivan, Yang, & Yang, 2012).

Assimilating to the American culture was challenging for the Hmong due to their cultural beliefs and lack of exposure to western healthcare. The Hmong believe illness has either spiritual or physical causes (Lor, Xiong, Park, Schwei, & Jacobs, 2016). They do not have an understanding of preventive medicine or of chronic disease, resulting in low screening rates and later stages of diagnoses and high cancer mortality rates (Lor, 2017).

The Hmong also encountered changes in their family dynamics while living in the US. Traditionally, the Hmong have a patriarchal society. Elders, defined as someone who is 50 and older, are seen as the “ones who possessed the wisdom and experiences necessary for survival of their family” (Frye, 1995, p. 273). Eighty-seven percent of Hmong elders have limited English proficiency, i.e., speaking English “less than very well” (National Asian Pacific Center on Aging, 2014). Consequently, there is a role reversal between children and elders, where children are seen as the ones who have the necessary experience for survival due to their English skills. Elders rely on children to translate and navigate the health care system (Lor & Bowers, 2017). In addition, the Hmong are excluded from national survey research due to language barriers (Lor, 2017).

Methods

Study Design

The findings reported in this paper are part of a larger mixed methods study (using survey, video, and qualitative interviews), which tested the feasibility of a data collection mode with older Hmong adults (Lor & Bowers, 2017) and requested that participants provided information about what they believed would encourage them and others from their community to take part in research. The Health Sciences Institutional Review Board (IRB) approved this study.

Data Collection Procedures

The first author, a bilingual and bicultural Hmong researcher, recruited participants from two senior community centers; she occasionally volunteered for the centers’ senior programs. During the senior programs, the researcher presented an oral description of the study including purpose, risks/benefits, and activities to Hmong older adults. Due to literacy challenges, potential participants were verbally instructed to provide their name and telephone number so the researcher could contact them to explain the study in more detail and screen for eligibility. The eligibility criteria were: (a) age 50 or older, (b) self-identified as Hmong, and (c) willing to be interviewed. If eligible, the researcher scheduled an interview at either the community center or the participant’s home, depending on the participant’s preference. Participants were also encouraged to identify other Hmong who might meet the criteria (snowball sampling).

Consents were obtained orally prior to the start of the interviews consistent with the Hmong people’s oral culture. The data collection process included two steps. Firstly, the participants and their family helpers completed the survey. Secondly, a follow-up interview was conducted to examine the feasibility of the survey data collection tool. This paper only focuses on one closed-ended question and two semi-structured, open-ended questions about Hmong participants’ willingness to participate in research (see Table 1). The entire interview process took 40 to 60 minutes and was audio recorded. Participants were compensated $20.

Table 1.

Closed and open-ended questions

One closed-ended question (survey)	“How likely are you be willing to participate in the research study if the researcher is Hmong?” using 5-Likert scale response categories of extremely unlikely, unlikely, neutral, likely, or extremely likely.
Two open-ended questions (qualitative interview)	For future research, does it matter if the researcher is Hmong or not? Would you still take part in their study? Please tell me more why or why not.” “What are things researchers who are Hmong or not Hmong can do to increase your willingness to participate in a research study?”

Data Analysis

The closed-ended question was analyzed using descriptive statistical analyses (e.g., proportions, frequency) in NCSS version 9.0. The two semi-structured, open-ended questions were translated by a group of bilingual and bicultural researchers from Hmong into English (Harkness, Pennell, & Schoua-Glusberg, 2004). Transcripts were analyzed using conventional content analysis (Graneheim & Lundman, 2004) of data obtained from the semi-structured interviews. First, four research team members read through the transcripts several times to achieve immersion in the data. Next, they individually coded the transcripts using labels that described themes common across all participants. For example, when the participant said, “...you know that person is someone you’ve known before. You know him/her and she/he knows you for a time. That person has helped you and you have helped him/her...” the statement was coded (1) “knowing the person” for the first response and (2) “reciprocity” for the second response. After all transcripts were coded, they were sorted into categories based on similarities and differences. For example, “knowing the person” and “reciprocity,” were subcategories of “insider characteristics.” This process of coding and categorizing was done iteratively to ensure conceptual saturation across themes (Morse, 1995). Saturation was reached on one central theme—trust, followed by conditions that influenced trust. Research members resolved disagreements or discrepancies in coding based on consensus by looking together at the transcript. There was a 98% agreement on the coding.

Results

Thirty older Hmong adults completed the study. The majority were female with an average age of 66 (see Table 2). Trust was a central influence on older Hmong people’s willingness to participate. They reported that the most obvious ways to achieve trust were to have a close relationship or to be “one of us”, to demonstrate reciprocity, and to have an absence of language barrier between the researcher and participant.

Table 2.

Demographic of Hmong Participants (N=30)

	N	%
Sex
Female	22	73
Male	8	27
Age mean	66
Marital Status
Married	18	60
Divorced	3	10
Widowed	7	23
Separated	1	3
Never been married	0	0
Income
Less than $10,000	13	13
$11,000 to $15,000	4	13
$16,000-$25,000	1	3
$26,000 to $50,000	0	0
missing	12	40
Number of Years of Education in US
0	21	75
1	3	10
2	1	4
4	1	4
7	2	4
Range of Number of Years Lived in US	9–36	23 (mean)

One of Us

Participants indicated that participation would be more likely with a researcher who was “one of us.” They linked those who were described as “one of us” based on how close their relationship was with the researcher. The closer their relationship, the more likely they were to consider the person to be “one of us,” the more likely they were to be trusted, and the more likely they were to participate in the research. Sixteen of 30 (53%) reported that if the researcher was Hmong, they would be “likely” to participate in the research. Likelihood of participation increases if one of the following conditions is met: the researcher is Hmong and the participant knows the researcher or the researcher is not Hmong but has a relationship with the participant.

Hmong and Knowing the Person

Participants shared that “knowing” the researcher, particularly if they are Hmong and speak Hmong fluently, would likely generate sufficient trust for the participant to be comfortable participating in a study. One participant shared, “… it has to be someone, who you trust a lot and know well then you are willing to” (Participant 2). They described having a higher level of comfort disclosing personal information when they knew and trusted the researcher because they would not worry about the shared information having a negative consequence. One participant shared “…Because that person is your person so if you want to speak more truthfully and harsh, there will not be any problems” (Participant 5).

As participants described knowing or not knowing the person, variations in the meaning of knowing the person were revealed. Participants described someone they knew as a person they were related to, who they had a relationship with, or with whom they experienced reciprocity. Participants defined reciprocity as an exchange that was mutually beneficial, requiring both parties to help each other in the Hmong community. An example of creating reciprocity involved helping the participant with a particular need (e.g., involved in medical decisions) particularly when the researcher’s expertise was useful. One participant shared, she would be more likely to participate if it was a person “who helped you and you had helped him/her…” (Participant 1). For some participants, the researcher’s role in the community was an important factor.

Fourteen (87.5%) of 16 Hmong participants also reported spontaneously that if they knew the researcher who was Hmong and related to them, they would be “extremely likely” to participate. Participant’s motivation and effort to participate in a study increased when they felt related to the researcher. “If the researcher is someone related to me, I want to participate even more” (Participant 4). One participant expressed, “…Someone who is related…this is something that might make you attend quicker” (Participant 13).

Although we did not ask participants about relationships with the researcher, several reported that having a personal history/relationship with the researcher, regardless of ethnicity (insider vs outsider), would increase their perceived likelihood of participating; in other words, someone they had interactions with occasionally or with whom they had built a friendship over time. A participant shared, “For me, Americans have come often, they come directly to my house…even if it is a Chinese woman, I am willing to, a Cambodian woman, I am willing” (Participant 25). Participants reported a willingness to participate if they had participated in research with the same researcher in the past and experienced no negative consequences. As a result, they trusted the researcher and believed the researcher would not misuse information they gathered. One person explained: “…if it is a person that I know or have participated in their research in the past and [I] know that they have not misuse my information then I am willing [to participate]” (Participant 5).

Not One of Us

Participants shared that those who were ‘not one of us’ were those who they did not know and with whom they had difficulty communicating due to language barriers. Even so, some participants described altruism as a motivation for participation with outsider researchers. For example, under the conditions that the researcher was not Hmong and the participants did not know the researcher, two (12.5%) of the 16 Hmong participants spontaneously stated they are willing to participate in research regardless of a Hmong or non-Hmong researcher because of a desire to help others: “…even if it is another race, I will help as long as it is a human being. Like I said, even with animals, we love [them] and if it’s humans, why not love [them]?” (Participant 20).

Not Hmong and Don’t Know the Person

Fourteen (47%) participants reported if the researcher was not Hmong, they were not willing to participate in research. One participant explained: “…if it is someone that I don’t know or if it was a stranger that comes and said, ‘I want to do this or bring someone with him/her,’ then I don’t trust them” (Participant 8). Thus, this participant perceived the researcher as an outsider.

Language Barriers

All participants reported language barriers was another determinant of their willingness to participate in a research study. Because 26 participants (87%) reported not being able to read English while 25 (83%) participants reported not being able to write English (Table 2), they shared that they do not trust researchers who do not speak their language. All participants anticipated language barriers with non-Hmong researchers and Hmong researchers who were not fluent in the Hmong language:

…so we can’t communicate, talking about important things like daily living, good and bad things, we wouldn’t be able to communicate. So there will be no place to move forward when we can’t understand each other (Participant 7).

Another participant shared, “If it is an American then I will not help because I don’t know. I can only talk to Hmong people” (Participant 24). These 26 participants reported they would not trust a researcher who they could not communicate with and who did not speak Hmong, including Hmong and non-Hmong researchers. In both cases, they saw these individuals as outsiders.

Hmong participants feared being lied to or that the researcher would misuse their information, which could lead to negative consequences for them, specifically losing their financial support from the government. One participant said, “There are Hmong people who lie. Other race/ethnicity also lie too” (Participant 9). One participant shared her perceptions of language barriers, saying, “You know living in this country, there are good and bad people. You don’t know the language so you are very scared” (Participant 16). Another participant explained the concern of loss of financial support:

They [referring to Hmong] are on financial assistance so if you come and talk to them, they may respond with something that is wrong and you guys publish it…that may cause problems with their assistance with food and financial. Like ‘they [referring to authorities] come to talk to you and you said this to them. However, you said this to us.’ So some refused to let you come because of this (Participant 16).

Similarly, another Hmong participant shared her concern about misuse of her data:

I am concerned because I am an elder and I do not know any written language, maybe, if there is someone that comes, they can come and change your documents or use it for something else that you do not know (Participant 21).

When asked whether they would be more willing to participate if the non-Hmong speaking researcher had an interpreter, 12 (40 %) participants said “yes”; 7 (23%) participants said “no”; and 11 (37 %) participants said “maybe.” Those who stated “no,” shared bad experiences with some interpreters in previous encounters, mainly in health encounters, where quality of interpretation was jeopardized, resulting in confusion between them and their providers. In addition, some participants reported that their goals were not met during the encounter, creating distrust. One participant explained,

The ones that are translating for you, there are some that only know Hmong and do not know English completely so when they take the state test, they don’t really pass. They don’t even pass and still try to translate (Participant 9).

A few participants also reported that even when a Hmong interpreter was provided, they still might experience distrust. This was especially true if the interpreter did not attend events in the community or had no reciprocity with participants; in that case, participants viewed the interpreter as an outsider. Thus, they would not participate in the research study, applying the same criteria to interpreters as they did to researchers. One participant shared “Even if there is an interpreter, the person is a stranger than I don’t really want to [participate]” (Participant 10).

Shift From ‘One of Us’ To ‘Not One of Us”

There are two conditions that could shift participants’ view of researchers who are “one of us” (insider status) to “not one of us” (outsider status). These include lack of community involvement and language barriers. When researchers who are considered “one of us” (insider status) fail to participate in community events, their status is left in doubt, potentially shifting them from an insider to outsider status. A participant shared, “They may be Hmong but they don’t attend community events like funerals” (Participant 9). As a result, he would not participate in the research study. Regarding language barriers, when Hmong researchers do not speak the Hmong language well, participants reported seeing them as outsiders despite their ethnicity. Consequently, not knowing the Hmong language creates a very low level of trust, decreasing participants’ willingness to participate in a research study. As one participant stated, “if the person is Hmong but does not speak Hmong, is the person really Hmong?” (Participant 3).

Strategies to Become One of Us

Despite the fact that most participants did not anticipate participating in research unless the researcher met the specific characteristics, some offered suggestions for increasing their trust, thereby increasing their willingness to participate in research. These strategies included calling out a connection and creating reciprocity.

Calling Out a Connection

Because trust was described as necessary for willingness to participate in research, some Hmong participants shared that if the researcher knew an active member of the Hmong community, particularly someone who had provided services or assistance to the participant at some time, attended community events, or knew the participants’ child or other relative, then the researcher can name that individual and call out their relationship. One participant shared,

Well if I have a friend, I can talk to them about [the researcher/research]... I can share with them [friends/elders] and if they like it then I will tell them and if they are willing to then I will tell her [researcher] (Participant 24).

Thus, the known community member acts as an intermediary, conferring trust (by proxy) in the researcher by calling out the connection.

Consistent with earlier descriptions of knowing the person, trust was described as even more likely, and stronger, if the known community member was a relative of the potential participant and if the known community member initiated the introduction between the researcher and potential participant. This would allow the researcher to establish a relationship (albeit proxy) with that potential participant, increasing participants’ willingness to participate in the research study.

Creating Reciprocity

Researchers could find opportunities to create reciprocity with potential participants to increase trust and develop a relationship. For instance, some participants described how the researcher provided them with health information when they had to make critical health decisions. In return, if the researcher wanted them to participate in a research study to learn about Hmong culture, the participants would be willing to participate in the study and teach them about their culture. One participant explained, “For example, if she or he helps teach us about health treatments for illness then if she or he wants to learn about our culture, we can teach them” (Participant 6). Through the process of reciprocity, participants shared that they would develop relationships with each other, changing the researchers’ identity from an outsider to insider.

Discussion

The study findings suggest that trust seems dependent on whether the Hmong participants see the researcher as “one of us” (insider) or not (outsider), thereby influencing their willingness to participate in a research study. This finding is consistent with findings from other studies about barriers to recruitment of racial/ethnic minority populations (Yancey et al., 2006).

“One of us” is comprised of several characteristics including knowing the researcher, having a relationship, having some reciprocity, attending community events, and speaking the Hmong language fluently. When all of these characteristics are present, participants’ level of trust is highest. Some characteristics mentioned by Hmong participants are consistent with those mentioned by Latinos; Latinos distrust researchers because they do not know the researcher, and language barriers also prevent them from participating in research (Brown, Marshall, Bower, Woodham, & Waheed, 2014). However, the conditions that influence trust in the Hmong participants are different from those identified in other studies (Yancey et al., 2006). For example, African Americans distrusted researchers because of fear of mistreatment and exploitation, which stemmed from the historical context of research done with the African American community (Yancey et al., 2006). In contrast, this study found that fears among Hmong participants were related to concerns of researchers misusing their personal information, resulting in a loss of financial resources.

Another condition that influenced trust in this study was a shared language between Hmong participants and the non-Hmong researcher or Hmong researcher. This finding is contrary to London, Hurtado-de-Mendoza, Song, Nagirimadugu, & Luta (2015)’s study with Spanish speaking participants which found that those who only spoke Spanish were six times more willing to participate in clinical research compared to those who spoke both Spanish and English. In contrast, interestingly, most Hmong participants preferred a Hmong speaking researcher as well as an interpreter, who they knew, who participated in community events, and/or had a relationship characterized by reciprocity. This is consistent with what insider researchers have reported to be one of the advantages of coming from the community—knowing the language (Bonner & Tolhurst, 2002).

In contrast, an apparent “outsider” could be considered “one of us” by calling out a connection and creating reciprocity. These were two strategies for moving along the continuum from outsider to an insider. An apparent insider can also move towards an outsider status if they do not attend community events or have a language barrier. These findings revealed that there is a continuum of definitions about when a researcher is considered an outsider and insider. Insider vs outsider status can shift based on criteria not generally acknowledged by researchers. In addition, the definition of an insider vs outsider is quite complex; there is no single characteristic for how one is determined to be an outsider. However, from participant statements, insider vs outsider status seemed to be inextricably tied to relationships, where direct relationships indicate a researcher is an insider and indirect relationships indicate a researcher is as outsider. It is also unclear when an insider becomes an outsider. For instance, there are some Hmong, those generally defined by researchers as insiders, and who might see themselves as insiders, who are perceived by at least some community members as outsiders. The study participants’ perceptions of insider vs outsider mirrors the experience of other insider researchers in the literature (Breen, 2007; Hodkinson, 2005; Innes, 2009). For instance, Innes (2009) shared that his “closeness” to the research participants assisted with increasing trust in him and research project.

Although our study findings and other reported research suggest that cultivating a relationship with participants over time can contribute to greater recruitment success (Yancey, Ortega, & Kumanyika, 2006), it may raise ethical questions. Because researchers are required by their university’s internal review board to be neutral and to maintain confidentiality, they are prohibited or discouraged from discussing personal information with participants. However, our study has identified strategies to increase trust that may challenge these approaches. To create trust, the researcher must develop a relationship that includes personal connection and reciprocity, and personal information is disclosed. This raises ethical questions: Are researchers doing harm to participants if they cultivate relationships (dual roles) with participants? And how can researchers ensure that they are collecting data in a manner that is not biased?

Implications

This study revealed that when using any personnel to conduct a study, whether it is a Hmong or non-Hmong researcher, trust plays a significant role in Hmong participants’ willingness to participate. Thus, it is important to consider whether or not the key researcher is someone who is well known, has a relationship with the participants, has a presence of reciprocity, and is an active member in his/her community. More research is needed to understand how these characteristics could be assessed in practice.

Interestingly, we discovered that language also influences trust for Hmong participants. Participants shared that having an interpreter doesn’t really overcome the language barrier if the interpreter does not: (a) know the participants, (b) attend community events, and (c) know the Hmong language well. Most importantly, when the interpreter does not speak the Hmong language well, it contributes to higher levels of distrust, decreasing Hmong participants’ willingness to participate in a research study. More research is needed understand how to best evaluate language competency/proficiency among Hmong interpreters since it is possible that a Hmong person may pass an interpreting exam but be unable to speak English or Hmong fluently, which contributes to the success of increasing research participation among Hmong participants.

Another key implication is that researchers assume that if they are from the population being studied, they are considered an insider. This may be an invalid assumption since study findings suggest insider status runs on a continuum. This finding has implications for researchers when thinking about working in their own community; specifically, what is their impact on the community? Furthermore, the strategies suggested by the Hmong participants illustrate that a non-Hmong researcher can move from outsider status to insider status. However, it will require a large amount of time and commitment in working with such population to have insider status.

Limitations of Study

This study has some limitations. First, because it was part of a larger study, the questions asked of participants may have not been comprehensive, including strategies that would apply more to non-Hmong vs Hmong researchers. Also we did not ask questions about whether or not they would be willing to participate in the research study if a Hmong researcher was a person who was involved in the community but did not speak the Hmong language well. Future studies could replicate this study with other Hmong people or other racial/ethnic minorities and ask more in-depth questions. In addition, because we interviewed non-English speaking Hmong, the results may not represent those who are English speaking racial/ethnic minorities. Future research could interview other English speaking populations. Also, this study was done with one group; it is unclear how much participants would disclose to an outsider vs insider. Future research could conduct a comparative study of a group with researchers the participants know and one they do not know and examine the level of participant disclosure and how willing they are to participate in research. Since the study sample was 73% female, it is unclear if having more men in this study would provide a different view. Finally, because the first author is an insider by many definitions, including being the same ethnicity as the participants and being well known as involved in the participants’ community, it is difficult to tease out whether or not the author’s identity had an impact on participants’ responses.

Conclusion

Recruiting racial/ethnic minorities into health research continues to be a problem for researchers. This study explored one racial/ethnic minority group, Hmong older adults, their perceptions of insider vs outsider researchers, and the impact of these perceptions on their willingness to participate in research. Our findings revealed that if Hmong participants know the researcher at an intimate level, have a personal relationship, and recognizes a sense of reciprocity between them and the researcher, then they are more willing to participate in a research study. These findings may have implications for researchers to building in time and commitment to increase trust among racial/ethnic minorities, thereby can increase research participation.

Figure 1.

Hmong Participants’ Language Proficiency Level