Recruiting and Retaining Arab Muslim Mothers and Children for Research

Abstract

Purpose

To describe successful and not-so-successful strategies for recruiting and retaining Arab Muslim immigrant women and their adolescent children for research.

Design and Methods

A longitudinal study of mother-child adjustment of Arab immigrants to the US is used for illustration. A panel of experts was assembled and provided culturally specific advice about gatekeepers, advertising, data collectors, data collection, and how to track and encourage participation at subsequent time points in the study.

Findings

Most of the strategies recommended by the panel were overwhelmingly positive, including advice about data collectors, how to collect data, financial incentives, avoiding offending families, and personal contacts. Hiring data collectors who were able to establish personal and culturally appropriate relationships with study participants was the single most successful recruitment and retention strategy. Advice from cultural experts about which gatekeepers to engage and how to advertise for study participants was not productive.

Conclusions

Researchers should not only assemble a panel of cultural experts to provide advice about group specific strategies to build trust and maintain cultural sensitivity, but also to budget generously for time for data collectors to build and maintain rapport with study populations who, like Arab immigrant women, highly value personal relationships.

Effective recruitment and retention of study participants is an essential component of generating valid research findings. Recruitment procedures must adequately sample the population of interest or else study findings are skewed and limited in their generalizability. Low retention rates further compromise the scientific integrity of the research by biasing study findings to sample characteristics that are related to ongoing study participation.

Ethnic minorities are one of many vulnerable populations that pose recruitment and retention challenges (Dutton et al., 2003). Barriers specific to recruiting and retaining ethnic minorities include language barriers, cultural differences, family and community gatekeepers, and mistrust about exploitation and ethnic stereotyping (Jaber, 2003; Keller, Gonzales, & Fleuriet, 2005; Levkoff & Sanchez, 2003). The literature on this topic indicates the importance of building trust and establishing ongoing, culturally sensitive relationships that are responsive to the needs and concerns of the group under study (Andrews, Felton, Weyers, & Heath, 2004; Carter-Edwards, Fisher, Vaugn, & Svetkey, 2002; Keller et al., 2005).

Using bicultural community workers and community linkages to build trust and maintain cultural sensitivity is a common approach for recruiting and retaining ethnic minorities in research studies (Andrews et al., 2004). Yet, many options exist for engaging and using community linkages. Some evidence indicates that strategies for using community linkages to build trust and maintain cultural sensitivity should be specifically tailored to the study population. For example, Harachi, Catalano, and Hawkins (1997) found that the most effective strategy to recruit and retain African American, Latino, Native American, and Pacific Islander study participants differed according to ethnic group. In Pacific Islander and Hispanic groups, the contact had to be by a person who was personally known by the study participant. In African American groups, direct contact with someone who was personally known was not necessary as long as the person held a respected role in the African American community. For Native American groups, the direct contact had to be by someone who was not only personally known, but also held an esteemed role, such as a respected elder.

In this paper we describe our experiences recruiting and retaining Arab Muslim immigrant women and their adolescent children for research in the US. This information should be useful to other researchers to avoid the time and costs of trying unsuccessful strategies with this study population. It also indicates a context for tailoring strategies from the broader literature on minority recruitment and retention to specific groups.

The Arab Mother-Child Adjustment Study

The Arab mother-child adjustment study is a 5-year, longitudinal study of how maternal and child stress, coping, and social support as well as parent-child relationship, socio-demographic risk, and maternal adjustment affect child adjustment during transition from early to middle to late adolescence. The study sample is mother-child dyads from Arab Muslim immigrant families who live in the metropolitan Detroit area and have been in the US less than 15 years. The study was restricted to Muslims because Islam is not a predominant religion in the US, and it requires vastly different expectations for women and adolescents than what is expected in the mainstream. These vastly different expectations pose another layer of complexity for immigrant adjustment, particularly during adolescence. Of note is that the study began after the 9–11 terrorist attacks on U.S. soil by Islamic extremists and the subsequent U.S. invasion of Iraq.

Bilingual research assistants collect data from the mothers and children at three times, each 18 months apart. Presently, we are in Time 2 data collection. At Time 1, the children were 12 to 15 years old. By Time 3, the children will be 15 to 18 years old. Mothers and children provide data separately, but during the same scheduled appointment. They are asked a battery of fixed-choice questions in their language of preference (Arabic or English). Most women select Arabic and most children select English. Data collection occurs in the study participants’ homes, unless the mother prefers an alternate location. The battery of measures can technically be completed in 1 hour with the mothers and 45 minutes with the children if no interruptions occur and study participants do not introduce additional material for discussion. Typically, however, interruptions do occur, such as the arrival of unexpected guests or distractions from other family members. Most mothers also elaborate when answering fixed choice questions by describing their life circumstances. More often than not, mothers also socialize by serving coffee and conversing about topics unrelated to the study. As a result, the average time to collect data from the mothers is 3 hours. At the conclusion of the data-collection appointment, each mother is paid $60.00 as a thank-you for her and her child’s time.

The Local Arab Population and Culture

Metropolitan Detroit has one of the largest Arab populations outside the Middle East (Schopmeyer, 2000; Zogby, 1998). The local Arab population is comprised of Muslims and Christians as well as different sects within Islam and Christianity. Arab Muslim immigrants are relatively more recent arrivals than are Arab Christian immigrants, and they live in dense, geographically bounded neighborhoods with other Arab Muslims. Arab Christians are a more longstanding group, and they live in more dispersed and less ethnically dense neighborhoods (Schopmeyer, 2000; Zogby, 1998). Both the local Arab Muslim and Arab Christian communities have many well-established ethnic businesses as well as political, social, and religious organizations. Some of these Arab ethnic organizations are not exclusive to Muslims or Christians. Although Arab immigrants come from over 15 different countries in the Middle East, the Arabic language and the Arab culture unites them. Arab Muslims and Arab Christians differ in some cultural practices and norms, however, because Islam strongly prescribes how Muslims should conduct themselves in their everyday lives (Carolan, Bagherinia, Juhari, Himelright, & Mouton-Sanders, 2000).

A major component of Islamic law pertains to the family, including its structure, values, and role responsibilities. In general, Islamic family law requires highly traditional gender roles whereby gender relations are based upon respect rather then equality. Men typically are the heads of households and, through this role, they control the family finances and dictate what behaviors and activities are permissible for their wives and children. Children are expected to obey parents. Women’s main obligations are to marry, maintain the home, care for children, and protect the honor of the family. Family honor includes segregation of the sexes, particularly modesty in women and not being alone with men who are not immediate family, as well as refraining from behaviors that are prohibited by Islam, such as pre- or extramarital sex or drinking alcohol (Abu-Lughod, 1986; Corolan et al., 2000; Esposito, 1982; Kulwicki, 2000). On the other hand, how Islamic family law is interpreted and embraced depends on a host of social, economic, and political factors that not only vary throughout the Middle East, but are also shaped by migration to a Western country (Mogahdam, 2004). In addition to variation in the generalities noted above for the so-called “typical” Arab Muslim family, many Arab Muslim women after immigration have assumed community leadership positions and entered the workforce (Kulwicki, 2003). In turn, these role changes affect gender and family relations.

Maximizing Recruitment and Retention

The first step to building trust and being culturally sensitive with our study population began when we were designing the Arab mother-child adjustment study. During the design phase, we obtained input from four consultants with insider knowledge on Arab culture and the local Arab Muslim population. Their cultural expertise was through being Arab immigrants themselves and from their experiences as longstanding health and social services providers to Arab Muslims. One of the coinvestigators was also an expert on Arab culture. This coinvestigator (A. Kulwicki) is an immigrant from an Arab country and has years of research experience with Arabs, both locally in the US and abroad in the Middle East. These consultants and the coinvestigator (hereafter collectively referred to as the cultural experts) made recommendations about which gatekeepers to engage, where and how to advertise for study participants, whom to hire as data collectors, how and what data to collect, and how to track and encourage participation in subsequent points of the longitudinal study.

As we progressed from the design to the implementation phase of the study, we gained more insight about which of the recommended recruitment and retention strategies were most effective and we made modifications accordingly.

Gatekeepers

Our cultural experts recommended seeking endorsement for the study from key community gatekeepers, specifically imams of local mosques and political activists for the Arab community. Imams, for example, are highly respected as spiritual leaders. The cultural experts identified four specific imams of local mosques and one political activist with a local and national reputation. When we were preparing the grant, the coinvestigator with cultural expertise on Arabs approached these community gatekeepers for letters of support. They were very difficult to reach but were enthusiastic about the study once we were able to speak with them. They provided letters of support stating that they agreed to publicly endorse the study and to let us advertise the study in their organizations. The political activist also agreed to be our spokesperson on local Arab television and radio stations. Most likely, the community gatekeepers were enthusiastic and cooperative because they personally knew the coinvestigator through past, positive interactions.

When the study was funded, the same coinvestigator sought appointments with the imams. Because of busy schedules, their availability was not as immediate as we had hoped for. The political activist whose support we had enlisted when we were preparing the grant was no longer available. Between the time we had submitted the grant and received funding, he had relocated to address more national concerns.

Once the coinvestigator was able to meet with imams, they expressed concern that our study might negatively stereotype Arab Muslims, particularly since U.S. anti-Muslim sentiment had increased because of recent political events. Regardless, they pledged full support and agreed to advertise the study in their mosques. In the meantime, we had started data collection and found, as explained below, that recruiting study participants without the support of community gatekeepers was not problematic.

The cultural experts also emphasized the importance of husbands as family gatekeepers. Some mothers wanted to seek their husbands’ permission before committing to participating. According to our institutional review board, informed consent from mothers and assent from children were sufficient for their participation. Even so, out of respect for cultural norms about men having authority over family matters, we fully accepted mothers’ wishes. We not only encouraged mothers to take the time to seek permission, but data collectors also offered to talk with husbands about potential concerns. Husbands were assured that they could withdraw consent at any time if they changed their minds about allowing wives and children to participate. The main concerns among husbands were about disclosing personal details about the family and, like the imams, that study findings might negatively stereotype Arab Muslims.

On the other hand, some mothers did not want their husbands to know that they were participating in the study. Under these circumstances, we developed data collection and tracking strategies to facilitate complying with wishes for secrecy. For example, periodic contact over the course of the study was by personal phone calls to the mother. We sought direction from the mothers in advance on how these phone calls could occur and what messages could be left.

Incentives

Both the principal investigator’s prior research experience with other immigrant groups and the cultural experts on Arab immigrants confirmed that relatively recent immigrants typically need additional income. Thus, we used financial remuneration as the major incentive for study participation by providing $60 at the completion of each data-collection appointment. In addition, the study was advertised as providing the opportunity to earn up to $180 by participating in all three phases of the study. Although this financial incentive was more than what is typically provided, our funding source (the National Institutes of Health) accepted our rationale for this level of financial remuneration. We were aware, however, that the financial incentive might encourage some mothers to participate in a given phase of the study more than once. We cautioned our data collectors about this possibility and developed a tracking system to detect duplicate participation. In fact, a few study participants did indeed participate in Time 1 data collection more than once, sometimes by substituting a different child and sometimes by engaging the same child to participate twice. These duplicate study participants were reviewed on a case-by-case basis and the majority of them were dropped from the study.

Altruism and wanting to be heard were additional reasons for study participation. Some mothers said they chose to participate because they wanted to help other immigrants and voice their concerns as women and mothers.

Advertising

Based on knowledge that Arab children typically obey parents’ wishes, we geared advertisements toward the mothers and relied on each individual mother to gain the cooperation of her child. To offset the potential bias of recruiting from any particular site, we planned to advertise verbally and in writing at a variety of sites, including mosques; agencies that provide wellness care and resettlement assistance; and local ethnic newspapers, television, and radio.

Once we were able to meet with some of the imams who had provided the initial letters of support, we began to advertise in writing at two of the local mosques and at four health and social service agencies. Written advertisements included a notice with tear-off flyers in Arabic to ask eligible study participants to call a local phone number and leave a message in Arabic. Although people were taking the flyers, not one person responded by calling the local number. Most likely, Arab women were hesitant to call and speak with someone who was unknown to them. Therefore, we aborted all further efforts to formally advertise the study.

In the meantime, our data collectors had been highly successful with recruiting study participants through direct, person-to-person, advertisements. Many of the data collectors were highly visible because they lived in the community and were employed locally as professionals in other jobs, primarily as translators, medical technicians, teachers, and social service providers. Although data collectors did not advertise or collect data at their work, they capitalized on their reputations as trusted professionals and advertised the study whenever they had informal opportunities during nonwork hours to interact with other Arab Muslims. Data collectors also asked mothers who were already enrolled in the study to advertise the study to their social networks and personally vouch for them as trustworthy. Data collectors gave their personal cell-phone numbers for interested study participants to contact them directly.

We decided to capitalize on the success of direct, person-to-person recruitment and consolidate advertising efforts by designating one research assistant as the recruiter and have her recruit for other data collectors. Although this research assistant was able to solicit interested study participants when she was a data collector, she was unable to recruit anyone when she was functioning as a recruiter for other data collectors. Presumably her lack of success as a recruiter was because mothers were more comfortable making a commitment directly to the person who would be collecting their data. We concluded that having the recruiter vouch for the trustworthiness of another person was not sufficient. This conclusion was reinforced by our cultural experts’ advice that personal individual relationships were highly important for Arab women. Ultimately, we returned to our very successful strategy of having each data collector advertise for and recruit her own study participants.

Data Collectors

To reduce language and cultural barriers, we planned at the outset to have data collected only by bilingual research assistants who lived or worked in the local Arab Muslim community. Being a woman in her thirties or older was an additional criterion. First, based on Islamic principles of gender segregation and modesty, we anticipated that mothers would not be comfortable being alone with, or answering personal questions posed by men (Kulwicki, 2003). Second, the age differential between the data collector and the children would be acceptable for collecting data from teen-age boys.

The cultural experts emphasized that hiring the “right” data collectors was highly important. All of our data collectors were personally recommended by at least one of the consultants who vouched for the data collector’s interpersonal skills and reputation as trustworthy. In addition, the coinvestigator, who also knew many people in the Arab community, spent considerable time obtaining additional, informal references to augment consultant recommendations. After they were hired, we periodically audited each of our data collectors by randomly selecting one of their recent study participants to assess mothers’ perceptions about how they and their children were treated. Our study participants appeared to enjoy the opportunity to share these perceptions and may have interpreted our efforts as another indication of respect.

To capitalize on the importance of personal social relationships for Arabs and to build trust and rapport as a strategy for retention, we tried to have the same data collectors assigned to the same mothers and children for all three time points in the longitudinal data collection. An unintended consequence of this retention strategy was that data collectors were highly motivated to retain study participants in their caseload because they had already expended considerable effort to recruit them. Nonetheless, we anticipated that not all data collectors would remain employed over the course of the 5-year study. The data collectors introduced this possibility to the mothers at the close of each data-collection appointment and obtained the mother’s permission beforehand whenever we wanted to assign a new data collector. Obtaining the mother’s permission beforehand was consistent with the lessons we learned from advertising the study: Having a third person vouch for the trustworthiness of another person was not sufficient. Rather, someone who was already known (e.g., the data collector) needed to personally vouch for the trustworthiness of a new person.

We expected attrition to be higher if the data collector changed. Thus, we tried to create positive work experiences for our data collectors. We paid them generously and were highly cognizant of their needs. Because the data collectors were professionals and mature women who were balancing the study with other employment and family commitments, they were allowed to autonomously manage their workload to suit their individual life circumstances. Their goal was to earn additional income during after-work hours (i.e., late afternoon, early evening, and weekends), which coincided with when the study population was accessible. The only stipulation was that data collectors were asked generally to collect data from at least one mother-child dyad per week, but we relaxed this expectation during religious holidays and if they had more pressing priorities.

We also made the training and exchange of data-collection materials as convenient as possible. Meetings and secure storage for exchanging data-collection materials were located in the data collectors’ community rather than at the university where the study was funded. A few times a year, however, meetings were held in the principal investigator’s home. These meetings included serving food prepared by the principal investigator. Although meeting in the principal investigator’s home was less convenient for the data collectors, it was a more relaxed atmosphere and allowed the principal investigator to offer hospitality in a manner that is highly valued in the Arab community (i.e., by cooking and serving home-made food to guests). The data collectors appeared to enjoy and appreciate the hospitality. Nonetheless, we were cognizant that meeting outside of their community could be perceived as burdensome if it occurred too frequently.

Data Collection

Before the study began the consultants reviewed the measures for relevance as well as for content that might be considered offensive to Arab Muslims. Potentially offensive items were removed so as to not dissuade study participants from continuing in the study or referring other potential study participants. For example, as previously stated, Islam prohibits alcohol and recreational drug use and pre- and extramarital sex. Transgressions against these norms bring shame on the entire family and asking if these transgressions occur could be construed as an insult to the family’s honor.

An author of two of the copyrighted study measures did not want his measures altered by dropping potentially offensive items about children’s sexual thoughts and behavior. Initially, we complied. Quickly, however, both the data collectors and study participants communicated their discomfort over these items and confirmed the consultants’ original concern that causing offense could limit study participation. All of the study measures were pilot tested after they were revised with 30 mother-child dyads. This pilot study indicated that the revisions did not adversely affect internal consistency reliability or concurrent validity.

The consultants also recommended collecting data in the study participants’ homes and having a conveniently located, alternative site in the community for mothers who might be reluctant to offer their homes. First, Arab Muslim women who are relatively recent immigrants either do not drive or do not have regular access to a car. Second, because of norms about gender segregation and modesty, Arab Muslim women are unlikely to take a taxi or use public transportation. For privacy and logistic reasons, we did not want to rely on husbands or other male family members for transportation.

The mothers generally preferred having the data collectors come to their homes unless they wanted to keep study participation from their husbands. In these instances, data collection typically occurred at a relative’s home or in the data collector’s home. Although optional space was available in a local community center, it was never needed. The mothers seemed to prefer the home setting and mothers and data collectors were able easily to arrange alternative homes for data collection.

Data-collection appointments were flexible and took into account that unexpected visitors could take precedence over already scheduled times. Data-collection appointments also included extra time to establish rapport and build long-term relationships. For instance, data collectors accommodated study participants who wanted to socialize or talk about issues that were important to them, even if the topics were tangential or not related to the study.

Maintaining Regular Contact and Tracking

To increase retention for the 5-year study, the data collectors made personal phone calls directly to the mothers every 6 months until the mothers and children completed the study. These phone calls were, in part, for data collectors to continue to build rapport and capitalize on the importance of personal relationships for Arabs, and in part to track changes in address. Relatively recent immigrants move often, for example, when their financial situations improve or they need more space for newly immigrating relatives. Another tracking strategy was to obtain the name and phone number of a contact person who would be unlikely to move but would know the whereabouts of the mother and child if they moved.

Phone contact for address confirmation rather than seeking confirmation though the mail was also based on advice from the cultural experts. Because of language barriers and cultural norms for men to mediate their wives’ relationships with the outside world, husbands often handle incoming mail. We purposely avoided contact by mail because some study participants did not want their husbands to learn of their study participation.

Many of the recruitment and retention strategies described above required keeping track of information about each mother, including her current address and phone number, her designated contact person, her specific data collector, her individualized preference for how phone contact could occur, and, when necessary, whether she granted permission for a new data collector to be assigned. A computerized tracking system was designed using the Microsoft ACCESS software program to specify and sort these fields of information, as well as time points for subsequent phone contact and data-collection appointments to facilitate scheduling these contacts.

Discussion

Most of these strategies to recruit and retain Arab Muslim immigrant women and children were overwhelmingly positive. We reached our target sample size of 607 mother-child dyads for Time 1 data collection 2 months early. Ongoing efforts to track study participants every 6 months show that mothers and children are looking forward to subsequent study participation. Initial success with Time 2 data collection confirms this belief and indicates that we should continue to be able to retain study participants. Of note, however, is that recruitment was slow during major Muslim holidays, such as Ramadan and Al-Adha, and over the summer months when both data collectors and study participants were more likely to return to their homelands to visit relatives. In contrast, some of the heaviest recruitment for this study occurred during Christian holidays, such as Christmas.

These positive results are noteworthy given that Arab Muslim immigrants in the US are a vulnerable population. They hold vastly different cultural beliefs from the majority population and perceive substantial discrimination from non-Muslims, particularly since the September 11th terrorist attacks and subsequent war. As a result, the Arab Muslim women and children in this study should have been more hesitant to participate in a study such as ours: a study that is not only explicitly about Arab Muslims but also includes personal questions about family matters.

Much of this success can be attributed to the preliminary advice provided by our cultural experts, including the coinvestigator. They accurately advised us about hiring data collectors who were reputable and culturally appropriate, as well as how to make employment highly attractive. The cultural experts were also correct about providing strong financial incentives for study participation, collecting data in participants’ homes, revising questionnaires to avoid potentially offending family reputations, and relying on phone contact for tracking.

On the other hand, advice from the cultural experts about the importance of community gatekeepers was not entirely practical or fruitful. The community gatekeepers provided letters of support when we were preparing the research proposal and were willing to publicly endorse the study once it began. Yet, they were difficult to reach and their lack of availability posed problems for our time schedule. Of note is that our time schedule, like most research studies, was dictated by when the grant was awarded, which could have not been fully known in advance. More importantly, the endorsement by community gatekeepers was not necessary or the most effective route for recruiting study participants. This finding is not consistent with Jaber’s (2003) success with using community gatekeeper to recruit Arab Americans for her study about diabetes, which is the only literature we could locate on recruiting immigrants in general or Arabs in particular. Perhaps we would have been less successful in our recruitment efforts if we had not received eventual endorsement from community gatekeepers, or worse, had community gatekeepers publicly voiced disapproval of the study.

Advice from the cultural experts on where and how to advertise the study was also not profitable. Not one of our study participants was recruited through the formal written advertisements that we posted in mosques or public agencies. Perhaps, however, formally advertising through written material may have sensitized potential study participants to be more responsive to data collectors when they were advertising directly through personal contact. In any event, personal contact was the only means of advertising that ultimately resulted in recruiting people into the study.

Disentangling the relative contribution of any one of our single recruitment and retention strategies is difficult. Clearly, personal contact by the data collectors was crucial to the success noted above and their efforts appear to have contributed the most. Yet, our consultants hand-picked the data collectors, and a coinvestigator devoted additional time to confirming the consultants’ recommendations. Personal contact by the data collectors may not have been effective if we hired the wrong people. In fact, the wrong people could have damaged the reputation of the study and severely curtailed recruitment. The financial incentives as well as the positive data collection and work environments most likely also contributed greatly to our recruiting and retaining both study participants and data collectors.

Other researchers employing the same strategies with Arab Muslims in other geographic locations might have more difficulty. Without question, the people in our study reported perceived discrimination based on their Arab Muslim identity. Even so, the size and political representation of the local Arab community most likely helped them feel relatively less discrimination than do Arab Muslims residing elsewhere in the US or in other non-Arab or non-Muslim countries. Greater perceived discrimination might cause Arab Muslims in other geographic locations to be less responsive to the efforts described in this paper to build trust and maintain cultural sensitivity.

Conclusions

These findings indicate that researchers should invest the time and effort to assemble a panel of cultural experts to provide advice about group specific strategies to build trust and cultural sensitivity with vulnerable populations. Having at least one coinvestigator who is an insider with the study population is an asset that cannot be overstated. In addition, researchers should budget generously for incentives for study participation and time for data collectors to build and maintain rapport with study populations who, like Arab immigrant women, highly value personal relationships.