#66594: Conducting Culturally Sensitive Psychosocial Research

Basic research has been defined as social research that is used for the primary purpose of advancing knowledge without the consideration of the economic or social benefits, practical ramifications to practice, or transferring the information to key individuals in sectors who can apply the findings [1,91]. It is a bottom-up strategy of collecting information to advance knowledge [156].

Traditionally, a significant amount of financial and time investments have been made in basic research, resting on three assumptions [5]:

These goals of basic research have been called into question, particularly during a time when greater fiscal accountability is required.

Applied research refers to using social research methods to learn about a particular issue (i.e., violence, depression) with specific practical outcomes in mind, with the assumption that specific groups or society as a whole will benefit from the research findings [1,136]. This has been referred to as "the science of the concrete," in which the goal is to study everyday practices to ensure that multiple perspectives are captured and not privileging any one voice [124]. The goal of applied research is the immediate application of research activities and outcomes into practice and disseminating research findings to professionals in order to implement the information [5,91,136,156,185]. Applied research, unlike basic research, is often conducted in real-life situations, in which it is not possible to control the settings or conditions under which the study is conducted [6]. There have been discussions of whether this dichotomy is artificial. Some have argued for a middle category of "use-inspired research," in which the goal is to maximize knowledge and application [91]. The term "implementation science" has also been used, which emphasizes application of the intervention or policy [185]. As with all dichotomies, it is presented in absolutes. For example, it is assumed that basic research is guided by theory while applied research does not. The reality is that applied questions can only be asked and answered within a theoretical context [92].

What is knowledge? What knowledge base(s) does a particular professional field draw from? What are the different sources of knowledge that guide decision making in clinical practice? What is known, and beyond this, how do clinicians know what they know? These might sound like philosophical and rhetorical questions, but in reality, exploration of these questions raises critical awareness and understanding of one's professional discipline, which will ultimately shape how practice and research are conducted [7].

Understanding the nature and sources of knowledge is known as epistemology [7]. It has been proposed that there are four sources of knowledge [8,9,10,11,12,13,93,125]:

The goal of this section is not to elevate one type of knowledge over another; it is important to weave the four types of knowledge in an integrated manner. Decision making in clinical practice, drawing from all four sources of knowledge, can decrease distortions and biases that might result if only one primary source of knowledge is used [14]. Multiple sources of knowledge serve as a platform for practitioners to check the accuracy of their decisions [14].

Because this course focuses on conducting psychosocial research with ethnic minorities, it is important to review the characteristics of empirical research. Characteristics that are unique to empirical knowledge that set it apart from the other types of knowledge have been identified [15,186]:

Empirical knowledge may be further characterized as descriptive, causal, procedural, and relational [94]:

As briefly noted, there are three classifications of research designs. These may also be considered three approaches to obtaining empirical knowledge [2,15,95,96,187,188]:

All research methods are guided by a research tradition or an epistemological paradigm. As noted, epistemology is the study of the roots of knowledge. An epistemological paradigm, or research tradition, offers a set of assumptions about how researchers view the problems to be studied, how to go about collecting data, and how to analyze and interpret the data [16,17]. Researchers are guided by a specific epistemological framework that informs their knowledge limitations, and the research methods allow researchers to discover what they believe can be known [18]. Consequently, every researcher has a unique epistemological perspective when using a specific research approach (i.e., quantitative, qualitative, mixed methods). Not taking into account one's epistemologic framework may introduce biases when collecting and analyzing data [126].

In this course, four research traditions will be reviewed: logical positivism, interpretivism, critical theory, and feminist theory.

There has been increased attention about the need for researchers to reflect and be transparent about their positionality. Positionality refers to the disclosure of the researcher's sex/gender, racial/ethnic, religion, gender orientation, and other self-identifications and how these social categorizations may affect the research process. Reflecting on their own background and experiences with privilege, marginalization, and oppression, researchers should deeply explore how these issues shape the formulation of the research questions and how data are collected and analyzed [196,197]. This is not an one-time reflection, but a sustained work of continual self-reflection of the dynamics of the researcher's multiple social positions and their influence on power and privilege in relation to research participants and the community [198,199]. Positionality mapping may be beneficial in starting the process of transparency and reflection [198].

The research process begins with asking culturally relevant research questions—questions that improve the quality of participants' lives and align with the norms and values of the group being studied [165]. It is important to focus on specific characteristics, experiences, or attributes that are relevant to the group (e.g., racial socialization, immigration, acculturation, enculturation, prejudice, bias) [166]. While simply comparing results to a White sample may be easier, it can inadvertently convey a deficit, give the impression of White supremacy, or suggest that one group is more well-adjusted or the norm [166,194]. Employing a strengths-based approach is recommended to avoid reinforcing negative stereotypes. For example, instead of examining the attrition and school drop-out patterns of African American youth, a strengths-based approach would to examine patterns of academic achievement in this group [166,194].

Use of appropriate theoretical frameworks should also be considered. Theoretical frameworks can generally be organized into three categories to help guide culturally sensitive research studies: generalizability, group differences, and multicultural perspective. A generalizeability approach assumes that the construct being studied is universal. Social learning theory is an example of this type of framework. Group differences approaches are more appropriate when the goal of the study is to understand how and why two groups differ. A multicultural approach focuses one group and specific attributes unique to that group [167].

Intersectionality theory should also be considered when conducting cross-cultural research. Crenshaw first coined intersectionality to mean that an individual's social identity often overlaps with other social categorizations, such as race, gender, class, age, ability, and socioeconomic status, all of which are connected to privilege and marginalization [188]. Intersectionality theory acknowledges the complexity of lived experiences, and multiple social categorizations are personified by an individual [195]. The convergence of these multiple forces inevitably shapes behaviors and experiences [200]. Intersectionality should be accounted for when formulating the research questions; designing the recruitment of research participants; collecting, analyzing, and interpreting data; and recommending practical implications and solutions [195]. Researchers often neglect to fully incorporate social context, power, and social justice despite purportedly indicating that their studies are grounded in intersectionality [200].

In the previous sections, various research traditions and how they guide the different classifications of research design were reviewed. As noted, research designs can be viewed as the overall plan by which empirical knowledge is acquired.

Quantitative research designs, guided by logical positivistic paradigms, assume there is one defined, objective reality that can be measured. Therefore, it is debatable whether these traditional research designs are optimum for conducting research with ethnic minority groups. If there is one universal reality, then should one assume that the monocultural and Eurocentric conceptualizations of a phenomenon apply to ethnic minority groups? Instead of accepting this theory, some assert that qualitative research designs guided by interpretivism may be more amenable when working with ethnic minority groups [36]. In this scenario, starting with the lived experience of research participants, researchers would not force a schema on the conceptualization of a problem. Furthermore, some assert that qualitative research fosters relationships between researchers and participants. This is particularly crucial in research with cultural groups and racial/ethnic minorities that are more relationship-centered [104]. Qualitative designs may also be more amenable to high context cultures, given that they tend to rely on oral traditions [105]. Finally, reflexivity, an inherent part of qualitative research, can identify the insidious ways in which Eurocentricism can enter into the research process [106].

Quantitative and qualitative designs are often viewed as mutually exclusive categories rather than principles that may complement each other [21]. Mixed method research designs fill this role, using the strengths of both quantitative and qualitative research designs to capture cultural nuances. Two studies involving African Americans and Chinese participants that used a mixed method design are discussed here. In these cases mixed method designs harnessed the strengths of both quantitative and qualitative methods. The cases illustrate the particular amenability of mixed designs to exploring the cultural nuances of a social phenomenon.

Researchers conducted both closed-ended surveys and ethnographic interviewing of African American and White children and their parents [37]. The researchers maintained that a mixed method design provides a layered understanding of the problem. The quantitative and qualitative analyses revealed interesting data about how society perceives and labels attention deficit hyperactivity disorder (ADHD) symptoms in ethnic minority children. African American girls with ADHD were more likely to be described as "bad," "stubborn," and "uncontrollable," while White girls were described as "reactive." African American boys were labeled as "endangered." Ultimately, these role perceptions influenced the types of interventions used. African American girls were given behavior modification interventions including spanking and disciplining, and African American boys were given more restrictions and behavioral modifications. In both cases, the ADHD symptoms were not viewed as indications of an illness but as behavioral problems. White girls received less professional help than their male counterparts because their symptoms were perceived as due to stressful events; boys' symptoms were believed to be related to genetics, which led more quickly to professional help. If the authors had used only one design method, these rich, multilayered explanations would not have been extrapolated.

Using a mixed method design as well, researchers explored factors contributing to the tendency for Chinese women to underuse cervical cancer screenings [38]. Qualitative methods, using focus groups with both screened and unscreened women, revealed that discomfort, modesty, and embarrassment impeded participants seeking cancer screening. In the quantitative research component, physicians were surveyed. This survey added to the data by identifying an additional cultural element, fatalism, which influences Chinese women's screening behaviors.

Participatory action research, or PAR, is premised on critical theory and feminist theories, and PAR researchers employ either quantitative, qualitative, or mixed-methods. PAR advocates that marginalized communities should take charge in developing the study, with the goal of instigating social change in their communities [201]. With this approach, the researcher does not enter the process as the expert; rather, the community and the research participants are the experts [201]. The community is engaged and participates actively in all stages of the research process. The research team collaborates with the community and stakeholders to identify a meaningful problem that they want to solve. Existing strengths, resources, and assets of the community are harnessed, with an emphasis on co-learning between the community and the research team [202]. Co-planning and communication with the community help to mitigate the fear and mistrust many communities have with researchers. Community engagement and buy-in are crucial in order to create long-lasting, feasible, and sustainable solutions [201,202]. Consequently, knowledge is co-created and shared by not only the researcher but by the researched. Ultimately, this process shifts the power dynamics in research projects [199].

One of the challenges of conducting culturally sensitive psychosocial research is defining the population to be studied. For example, a researcher may be interested in studying Hispanic families. The term "Hispanic" refers to a large and heterogeneous group who share some threads of similar cultural characteristics, values, and belief systems. However, to categorize all individuals from Latin America as Hispanic minimizes the tremendous diversity of individuals born into one of the several nations that make up Spanish-speaking Latin America [40]. A research participant who is Mexican American may reflect various demographic, social, political, and educational differences compared to a Cuban American participant [40]. To classify Chinese from Taiwan or from Cambodia or Chinese Americans from the United States all into one group labeled "Chinese" would undercut the political history affecting each of these Chinese subgroups. Chinese individuals from Cambodia, for example, who were targeted by the Khmer Rouge between 1975 and 1979 and experienced torture, political persecution, and harsh treatment because of their ethnicity, have starkly different social realities than those of third-generation Chinese Americans born in the United States [41]. Similarly, when studying Muslims, it is vital to remember that this is a diverse group, with differences in religious beliefs/traditions, national origin (e.g., Pakistan, Bangladesh), and ways of self-identifying [107]. A classification like "non-White" may be viewed as derogatory as it is not a racial/ethnic category. It is analogous to using "non-male" as an identifier of sex [166].

Many times, researchers employ demographic variables or self-identified racial and ethnic categories to capture the concept of culture [40,42]. For example, research participants might be asked to designate an ethnic category or might be asked where they were born. These demographic variables are then used to define culture. The simple demographic variable of birthplace does not necessarily capture the subjective nature of culture, which is comprised of beliefs, norms, and values, nor does it fully capture the dynamic and fluid nature of culture [43,44]. Even asking seemingly basic demographic questions (e.g., marital status) may require cultural nuances. For example, some cultures have a tradition of long engagement periods, and these individuals may not feel comfortable marking "single" or "married" for their marital status [107]. In other words, demographic variables may serve as proxies for culture, but they do not have inherent psychological, cultural, and/or social meanings [194].

Researchers often define ethnicity based on research participants' self-classification. The rationale for using self-identification classifications warrants merit. It can be argued that individuals know best what their cultural values and preferences are and how these influence the construction of their identity. Yet, how other groups ascribe meaning and identity also plays a role in the construction of ethnic identity [44]. Researchers describe a focus group study where the eligibility criterion was "self-identify as Chinese" [41]. One focus group participant maintained that she was Taiwanese, although she was ethnically Chinese. However, she did not want to be affiliated with the political ideologies of mainland China. Hence, her perceived ethnic and cultural identity was Taiwanese.

Best practices when assessing participants' racial/ethnic categorization include [166]:

At first glance, defining the population of interest does not appear complicated. However, one of the dimensions of culturally sensitive sampling requires an understanding of how cultural cognitive scripts and the tremendous within-group differences influence the definition of racial, ethnic, and cultural categories. Unfortunately, it is not simple to capture the dynamic nature of these scripts and their effect on identity.

There are many different sampling strategies that may be used to recruit ethnic minorities for participation in psychosocial research. As discussed, probability sampling designs pose challenges. Consequently, nonprobability sampling designs are used more frequently. Unlike probability or random sampling, nonprobability sampling designs are not concerned with obtaining a sample that represents the larger population of interest [2]. Convenience or purposive sampling, a type of nonprobability sampling design, is used to recruit participants who fit a predefined criterion, using multiple recruitment methods, such as disseminating information about the research study through flyers, advertisements, and word of mouth. Potential participants would then be screened to determine if they meet the eligibility criteria. Studies using convenience sampling include [45,46,47,48,141,169,203]:

Other researchers describe recruiting African Americans through Black churches' rosters for survey studies on diet, hypertension, and blood pressure [49,108]. Media outlets that target a specific group or are presented in the target audience's native language can also be helpful [108].

In addition to being creative with recruitment methods, alliances with community leaders are vital to success. In one study, researchers recruited Latinas using convenience sampling, offering financial incentives, childcare, and transportation, but they struggled with successfully recruiting for more than one year [109]. However, after the researchers formed an alliance with a well-respected pastor, who endorsed the study, and moved the data collection to the church, participants were much more likely to participate. One motivating factor was the women's interest in sharing financial incentives with the church. Research indicates that going into the community to develop relationships and establish trust/credibility is one of the most effective strategies for recruitment of African American research participants [168].

Personal referrals, including potential participants' close friends or family members or trusted individuals from the community, can be very useful [141,203]. In a longitudinal study with Latino immigrant families, researchers and interviewers recruited from school and holiday events. Before attending these events, however, the researchers asked mothers who lived in the community to accompany them. These women allowed the researcher to more easily enter into the group and be trusted by potential participants [142].

One major limitation of convenience sampling is that the research participants recruited do not necessarily represent the population of interest; they may be more educated and/or from higher socioeconomic brackets. Snowball sampling is another form of nonprobability sampling whereby potential contacts are obtained through referrals from one or a few participants, and then referrals from those participants, and so forth, generating a "snowball" effect [2; 196]. Several challenges are inherent with this strategy. Some people may be hesitant to provide names and contact information for friends, family members, or acquaintances. Others may only have a name without specific contact information. This causes researchers to have to track down the information and create a system to keep track of the names obtained [50]. Additionally, participants will often give names from their own social network. Consequently, the individuals referred will be more similar to the initial set of participants, limiting the range of diversity of the sample [51]. Nevertheless, Rankin and Bhopal maintain that snowball sampling is an effective strategy, particularly when interviewers who ask for referrals are known and are familiar with the nuances of the community [50].

Another type of nonprobability sampling method is selection by ethnic surnames. The underlying assumption in using ethnic surnames as an identifier is that the ethnic group under study has surnames that are unique from other groups and are representative [52]. Himmelfarb, Loar, and Mott, for example, identified 35 Jewish surnames and used a random sampling of persons with these distinct Jewish surnames [53]. They found there were no major differences between samples obtained based on these surnames and sampling from lists obtained from Jewish organizations. Similarly, Shin and Yu compared the proportion of persons with the surname Kim in eight groups from different regional, occupational, and socioeconomic backgrounds [54]. They approximated that the surname Kim reflected more than one-fifth (22%) of the Korean population. Consequently, using Kim as the identifier was a viable method for selecting potential Korean American participants. In another example, Chinese elderly were identified using Chinese surnames from 1990 Medicare enrollment files and then compared to their demographic profiles from 1990 U.S. Census data [52]. The demographic profiles from these two datasets were comparable, supporting the use of ethnic surnames as a viable sampling technique. However, several caveats should be acknowledged when ethnic surnames are used for sampling purposes. A particular ethnic surname is not always exclusively representative of that ethnic group, and not all ethnic minority households may have a unique surname [55]. Finally, this technique does not identify those ethnic minority women who marry outside their ethnic group [52].

A probability sampling strategy called time-space sampling has been employed to identify hidden or difficult to access populations. Time-space sampling was developed by the Centers for Disease Control and Prevention (CDC) and has been used primarily for research with gay and lesbian groups. However, the concept can be extrapolated for research with ethnic minorities. Time-space sampling involves recruiting participants from the target study population during times and places where they congregate rather than where they live [56]. It involves three steps [56]:

There is no one preferred culturally sensitive sampling strategy that will ensure enhanced research participation and retention of ethnic minority research participants. Often, ethnic minorities are particularly wary of researchers asking for private personal information that is perceived to potentially place the group or community at risk [40]. Consequently, it is crucial for researchers to continually reflect how they can establish collaborative and trusting relationships with gatekeepers, community leaders, and research participants.

One interesting and collaborative exercise to promote engagement and collaboration is to conduct a transect walk. This involves the research team, community leaders, residents, gatekeepers, and other trusted individuals walking together through a defined path in the community. During this walk, the parties discuss the steps necessary to involve all interested parties and to ensure the research is culturally relevant and sensitive [126].

Concepts or constructs that are examined in a research study may not always be relevant for each cultural group. For example, the concept of "ethnic identity" for ethnic minorities who have been residing in the United States for many generations may not necessarily be as relevant or important as it is for recent immigrants [42]. Sense of self as a construct is also culturally scripted. Some groups, such as Hispanics and Asians, are collectivistically oriented; thus, their sense of self revolves around the group. Groups from Western industrialized countries generally perceive their positions as more individualistic [57]. This is often reflected in language. In Chinese culture, the concept of "self-esteem" is not equivalent to Western definitions. The closest (but not equivalent) concept is self-respect [170].

When asking about marital status as a demographic question, it is now commonplace to include a response allowing for cohabitation or a permanent relationship with a partner outside of marriage. However, in some cultures, there is no simple term to capture this concept [143].

In addition, future-oriented questions may be religiously dissonant to some research participants. In one study, Turkish and Moroccan participants were asked whether they believed their health would worsen, a question that confused and distressed them [110]. The authors speculated that the distress was in response to religious beliefs about fate; the participants did not believe they could comment about the future, as fate had already predetermined events. Conceptual or construct equivalence of an instrument is not easy to establish. Searching the literature is recommended to determine what has been written about similar topics and its relation to the target culture [58]. Consulting with experts about the topic and the culture of interest and conducting focus groups with individuals who are representative of the target culture may be helpful.

Scale equivalence refers to the extent to which the response choices have similar meanings across cultural groups [42,59]. The Likert scale is a common response format. Likert scales are a scaling method that measures the level of agreement or disagreement. Research participants might be asked to select from a 5-point Likert scale, in which 1 represents "Strongly Agree," and 5 means "Strongly Disagree." The issue of scale equivalence is whether these response options might be culturally influenced.

Researchers found that when an odd number of response options are used, Asian and Asian American research respondents are more likely to select the neutral response category [60]. For this ethnic group, selecting a neutral response conveys modesty; conversely, selecting an extreme response on a Likert scale would mean that one was boisterous [61]. This is culturally consistent with values in Asian groups emphasizing harmony and conflict-free relationships [60]. However, the converse has been found to be true for Hispanics. Hispanics correlated those who select a middle of the road or neutral response as either hiding something or responding insincerely [61]. One study found that Hispanics were more likely to select the extreme scaled responses compared to their non-Hispanic counterparts [61]. This is known as an extreme response style. It is the tendency for some research participants to select the extreme endpoints of a scale when responding to a question [62].

Acquiescence bias is another form of response style, whereby research respondents agree with the question(s) regardless of the content [62]. Some researchers speculate that cultural values emphasizing harmony influence this response pattern, and they posit that a cultural script in Hispanic culture known as simpatia, which emphasizes harmonious, cooperative, and positive social relations, has a role in acquiescence bias in this population [40]. In one study, Latino research participants who were less acculturated or had stronger Latino cultural orientations displayed higher levels of acquiescence bias [204]. In addition, when the interview was conducted in Spanish, this influenced acquiescence, even after controlling for acculturation and other participant characteristics. Other studies have found that certain racial/ethnic groups are more likely to choose the option "yes" in yes/no questions [166].

Religious beliefs may also influence how participants respond to usually closed-ended questions. In a study of Turkish and Moroccan participants answering questions about health and well-being, the individuals responded Al-ḥamdu lillāh, or "thanks be to God," instead of "yes" [110].

Courtesy bias is a similar term that refers to East Asian populations who tend to exhibit acquiescence bias [62]. A study of more than 20,000 employee surveys representing 19 different countries from Europe, Asia, Australia, North America, and South America examined the extent of extreme response styles and acquiescence bias in respondents from these different countries [62]. The findings indicate that countries that were more individualistic-oriented were less likely to provide acquiescent responses, with respondents expressing stronger opinions. Male-centered cultures also tend to emphasize assertiveness and, consequently, extreme responses. This study contributes to a body of knowledge related to the role of culture in data collection.

Operational equivalence refers to whether the procedures, ways of administering the instrument, instructions, and formatting of the instrument are logical for other cultures [144]. Many instruments, for example, have a time reference for the participant to consider (i.e., in the last week). However, not all cultural groups will share the same concept of chronology as the group for which the instrument was originally developed and targeted [58]. Researchers should also consider whether the mode of administration is culturally consistent with the norms of the cultural group. For example, some cultures may find it culturally offensive to have a young interviewer ask elder research participants personal questions [41,58].

Linguistic equivalence refers to the extent to which the content and grammar of question items have the same meaning when used for individuals in different cultural groups [42,58]. For example, the term "mental" can be misinterpreted as offensive by other cultures. In the Pakistani language, the term "mental" is negative, referring to a lunatic or a person who must be hospitalized [63]. The Khmer term for mental illness is ckuot, which also means crazy [41]. When the meanings used in an instrument are altered and ultimately interpreted as offensive or humorous in another cultural context, the reflective meaning changes [58].

In some cases, terms that exist in one culture do not exist in another. For example, the term "self-esteem" as defined in Western culture does not exist in Chinese language or culture [57]. The sense of self in Chinese culture is not individually rooted but oriented in relationships to others, and children are socialized to minimize their own roles in their accomplishments and to consider modesty, humility, and the feeling of others [57,170]. So, the concept of self-esteem is not easily translated.

Level of education and social class should also be taken into account when assessing an instrument's linguistic equivalence. For example, in Puerto Rico, the word "hair" is translated into Spanish as cabello. However, if the study involves Puerto Rican children this may not be the most recognizable term. The average Puerto Rican utilizes the term pelo; cabello is a more formal term among elder Puerto Ricans [111].

Dynamic equivalence refers to the translated instrument having a similar or equivalent effect compared to the original instrument [111]. In order to achieve dynamic equivalence, an instrument's items usually cannot be directly translated. For example, the term "no smoking" can be translated to défence de fumer in French; however, défence de fumer literally means "interdiction to smoke" [111]. The reader would likely eventually determine that this means smoking is prohibited, but the effect is not necessarily the same.

When conducting research with ethnic minority populations, it is frequently necessary to translate instruments into the target language. This can be particularly challenging if the target and source languages are different in terms of semantics and syntax (e.g., Chinese and English). The complexity is augmented by the fact that the cultural norms may be at different ends of the continuum (collectivistic vs. individualistic) [170]. There are three primary reasons for translating an instrument: it is more cost-effective than developing a new instrument, it increases the validity and reliability of the data, and it enhances the comparability of data across groups [145]. The specific translation procedures should be carefully selected in order to promote conceptual and linguistic equivalence of the instruments used across groups. Researchers should be familiar with the merits and limitations of the three common translation techniques [40,64,65].

As noted, ethnic minority communities may be mistrustful of researchers entering and potentially exploiting research participants. Historically, there have been cases of researchers breaching social and cultural etiquettes, not involving the community and their leaders in the research process, and abandoning the study group once the data are collected. Furthermore, when interpreted without a full historical and cultural context, data can pathologize communities and their members. It is for these reasons that many ethnic minority communities tend to be reluctant to participate in research.

Given this context, it is important for researchers and sponsoring organizations to develop relationships with communities and their leaders and to establish legitimacy and credibility. This involves the community's views regarding the level of authority and trustworthiness of the researcher and the sponsoring agency of the study. Researchers have identified two types of credibility: ascribed credibility and achieved credibility [69].

Ascribed credibility refers to a perceived attribute connoting position and authority. These attributes or characteristics might include age, gender, credentials, or race [41]. For example, in Asian culture there is an emphasis on hierarchal relationships. The age and gender of an interviewer might affect perceived credibility. In other situations, research participants may ask interviewers personal questions about their age, marital status, and education in order to assess the interviewer's ascribed credibility. Researchers should carefully train interviewers how to respond to personal questions and limit the amount of information they disclose [41].

Another way of obtaining ascribed credibility is by highlighting the role of the sponsoring organization or institution of the study. In some cases, studies may be imbued with a certain amount of ascribed credibility due to the affiliated institutions [41]. In one example, studies were sponsored by two well-known universities and a local school district. Because of the amount of respect attached to education and participants' own aspirations to have their children and relatives attend those universities, it was believed that a certain amount of ascribed credibility and legitimacy were obtained. In both cases, letters on the sponsoring university and school district letterhead were sent out introducing the study. In other studies with the different Asian American groups, they indicated that having an endorsement of the study from a respected individual from the community or a family member would facilitate research participation [113].

Achieved credibility is based on a researcher's ability to gain the trust of research participants [69]. How well interviewers are able to make participants feel comfortable, the level of professionalism conveyed by the different members of the research team, and the ability of the research team to involve community leaders and cultural experts all assist in lending legitimacy to research studies. For example, in the African-Caribbean community, researchers are often scrutinized to determine how sensitive and/or respectful they are of cultural norms, and it is in this way that credibility is achieved [106].

Researchers should be cognizant of the study group's communication style and incorporate this preference in the research process [146]. For example, in groups partial to oral communication making face-to-face contact or phone calls may be more effective than written communications (e.g., e-mail, letters, flyers). It has been observed that the Chinese language is not direct [203]. In Chinese conversation, "beating around the bush" is a sign of politeness, especially when interacting with those perceived to be in authority or in respected positions or with strangers. Therefore, it is important to know when to continue to question with gentle probes [203].

Giving is another way to obtain achieved credibility [70,141]. Giving, by providing incentives for research participation, produces a perceived equitable exchange between the researcher and the research participant [71]. Incentives can be either monetary or nonmonetary in nature. When conducting research with ethnic minority groups, it is crucial to take into account social context and cultural beliefs about incentives and how these beliefs may influence decisions to participate [114]. In one study in the Khmer community, community informants advised the researchers that monetary incentives may not be culturally appropriate because participants may not be comfortable in accepting checks [41]. Many of the Khmer locals worked only for cash. In this case, grocery coupons were offered, which was more culturally sensitive. Community stakeholders and other trusted community members and experts should be consulted [174,175].

Whether interviewers and participants should be matched based on race or ethnicity (racial/ethnic concordance) in order to help increase response rates is another issue that has been raised. Becerra and Zambrana argue that response rates can be enhanced by using same-race or same-ethnicity interviewers [36]. The premise of this perspective is that two individuals from the same group are more likely to share similar cultural values and therefore are more likely to develop rapport and communicate more easily [72]. There is a body of knowledge in the counseling field, particularly in counseling Asian American and immigrant clients, that ethnic matching decreases likelihood of early termination. The ethnic responsiveness hypothesis argues that there are beneficial outcomes when matching clients and counselors along ethnicity, language, and gender [73]. In one review of barriers and facilitators to racial and ethnic minorities' participation in research studies, researchers found that cultural congruence between the researcher/interviewer and the participant was key to promoting participation [113]. African American research participants liked having a "personal touch," and Filipino participants were more comfortable if they knew members of the research staff. In another study, African American participants preferred to have a racially congruent interviewer when the study involved racial content [126,147].

However, not all studies have found significant positive outcomes when matching interviewers and research participants [74]. In a study of infertility in the South Asian community, it was assumed that South Asian female research members would be more likely to connect with South Asian women with infertility issues because of their "insider status" [115]. However, in this case, the White female research members had easier access to participants because their "Whiteness" was ascribed to their being "experts" [115]. In another study, researchers tested whether ethnically matched interviewers and research participants influenced refusal rates [72]. Their findings did not lend support to the premise that ethnically matched pairs of interviewers and research participants affected refusal rates. In this study, the mental health diagnosis of the participant played a more influential role. A 2020 study found that attrition was higher among African American and Hispanic participants in racially/ethnically concordant dyads compared with discordant dyads [176]. The authors speculated that perhaps racially/ethnically matched researchers defied stereotypical assumptions of what participants believed scientists should look like [176]. Ultimately, it is vital to remember that shared ethnicity between the researcher/interviewer and participant does not necessarily guarantee trust and entry into the community [174]. Consultation with community leaders and cultural experts in determining the best approach within a particular community may be advisable.

It is recommended that interviewers should be fluent in both the dominant language and the language of the participant, and that research participants should be given a choice in deciding the language they prefer to have the interview conducted in. Interviewers may be required to switch between English and the respondent's native language depending upon the situation. This was the case in a study with Latino research participants in which participants asked to have the interview conducted in English but then switched to Spanish midway [204]. In another study, some participants expressed concerns with language barriers and started with Mandarin but then switched to English during the interview [203]. Therefore, interviewers should have both language versions available to make sure that the questions are asked in the same way to each participant and that on-the-spot translations are not being done.

A study of Asian Indian women in the United Kingdom demonstrated the extreme complexities involved when various languages must be accounted for when attempting to ensure an inclusive sample [75]. In this case, different interpreters fluent in a total of 45 languages had to be hired. Interpreters and interviewers were paired during the recruitment stage; however, it was difficult to determine which languages were needed. The authors describe how, after knocking at the doors of households, interpreters and interviewers would gesture in an attempt to communicate; family members were brought to the door to help; and in some cases, relatives were telephoned to help translate over the telephone. In conducting research studies with ethnic minorities, academic research training held by researchers may be thrown to the wayside when there is a commitment to conducting culturally relevant research that encompasses an inclusive sample [76].

The hiring of individuals to conduct the interviews is extremely crucial, as they are often the first point of contact to represent the research study. How potential participants are approached influences whether they decide to participate in a study [116]. Training for interviewers and recruiters should include sensitivity, flexibility, and adaptability [116]. Not only should the potential participant be approached with respect, but potential participants and their family members should be approached in a manner that minimizes power differentials [116]. In certain cultures, the research participant may seek permission or approval from family members before agreeing to be involved. Recruiters should always ask potential participants if they are comfortable in the setting [116].

Hiring interviewers who are ethnically matched to the target population or who can speak the same language as the research sample is ideal but will not solve all problems. Some have found that participants are less inhibited to talk with an interviewer who is racially/ethnically matched [166]. A study with low-income African Americans with mental health problems relied upon hiring interviewers who had clinical experience in working with this population [72]. Research studies focusing on sensitive topics, such as rape, may indicate a need for clinical interviewers who are female and have experience in dealing with trauma. Similarly, in a study about cervical cancer screening with South Asian women, researchers found that even having a male staff member accompany a female interviewer was perceived as embarrassing to the participants [148]. However, a study examining ethnic minorities' attitudes toward computer technology is more benign and may not necessarily require interviewers who have a clinical background.

Some recommend that intensive and consistent training and/or supervision sessions be provided to interviewers throughout the research process [72,177]. They also encourage reviewing specific topics with interviewers, such as: how to approach research participants; how to explain concepts about informed consent, confidentiality, and voluntary participation; the nature of the study; and the type of survey or interview instrument(s). When contacting participants and their families, research staff should remember that research participation may not be individuals' first priority, and the research should conform to the needs of the participant, being as flexible as possible, rather than the other way around [142,177]. For example, in many cultures, women's priorities revolve around their families and children, and data collection will take into considerations the needs of the entire family and family/childcare activities [148].

Training protocols should also be developed to establish how interviewers should respond to domestic violence, child abuse, elder abuse, and instances in which participants indicate they may hurt themselves or someone else. Although informed consent forms might emphasize the confidentiality of participants' responses, the limits of confidentiality should be defined, as should interviewers' roles and responsibilities in reporting cases of abuse. In some cases, interviewers, particularly if they are from the same community as the research participants, may perceive contacting legal authorities as a sign of betrayal to the community [41]. However, it is important to note that each state has laws that mandate the reporting of child abuse and neglect for specific professionals [77].

Monitoring interviewers and providing regular supervision is another important component of the ongoing training [41]. Having consistent and continual oversight and team meetings is essential to ensuring cultural sensitivity, research fidelity, and study coordination [177]. This should involve continually assessing and processing successes/failures and experiences of interviewers, research assistants, and other members of the research team [178]. In addition, the various feelings that are engendered when interviewing respondents should be monitored, particularly sensitive research topics such as trauma, violence, bereavement, or chronic illness. Attention should be paid to interviewers' potential to experience vicarious or secondary traumatic stress.

Safety protocols should be discussed and reviewed, with an emphasis placed on minimizing risks to the interviewer. Interviewers should feel supported by the researchers if they feel they cannot conduct an interview at night in neighborhoods with high crime and violence or if apartments do not appear safe. Specific guidelines and protocols should be established and followed for all interviewers and researchers who work in the field should carry their cell phones, fully charged [149]. This can include ensuring access to cell phones, alerting others as to plans and whereabouts, and scheduled check-ins and follow-up.

Questionnaires completed by interviewers should be reviewed and edited regularly. When question items are not asked, interviewers can be asked to contact respondents. However, this should not be the case when a participant refuses to answer a question. In such cases, the interviewer should clearly indicate that the participant refused to answer a question item. Missing data can be reduced in this way, and retraining of the interviewers can be conducted as necessary.

The interviewers' recruitment of respondents for the study should be monitored. When it is difficult to recruit respondents, interviewers may feel frustrated, resulting in a high attrition rate.

Best practices for culturally sensitive data checking and analysis include but are not limited to [155,166,179,206]: