Given the hard-to-reach nature of the intended sample, the authors utilized a snowball sampling strategy that commenced from their own networks in the United States to recruit participants. Specifically, participants were recruited by sharing the online Qualtrics survey link through American students and professional organizations’ email listservs, virtual messaging groups, social media, and word of mouth. Data collection began in October 2020 and ended in February 2021. After consenting to participate, individuals were able to complete the survey at their convenience. Participants who indicated that they were from a faith group other than Islam or reported being older than 29 were excluded from the study. To increase response rates, participants had the option to submit their emails to an external link at the end of the survey to enter a raffle drawing to win a $20 gift card.

One hundred eighty-two Muslim American adults (58 males, 124 females, M = 22.90 years, SD = 3.15, range 18–29) completed the survey and were included in this sample. Most of the sample reported their ethnic or racial identity as Middle Eastern or North African (n = 77), followed by South Asian (n = 55), Black or African American (n = 21), East or Southeast Asian (n = 16), White (n = 6), multi-racial (n = 6), and Hispanic or Latino (n = 2). Participants reported their socioeconomic standing using the MacArthur Scale of Subjective Social Status (Adler et al., 2000), a one-item subjective 10-point Likert scale, indicating where they stand regarding income, education, and occupational standing compared to others in the United States, where a larger score indicated a higher socioeconomic status (M = 6.59, SD = 1.58).

The 8-item centrality subscale from the Multidimensional Model of Black Identity (MMBI) (Sellers, 2013) was adapted to assess national and religious identity centrality. This subscale measures the importance a person places on their identity on a Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree), where higher scores translate to higher centrality. Two versions were administered to participants. The first version prompted participants to think about their American identity and the second version prompted participants to think about their Muslim identity. For the respective adapted versions, example items included: “Being American is an important reflection of who I am” and “In general, being part of my religious group is an important part of my self-image.” Total scores were calculated by summing the values of each item and the ranged was 8–56. In this study, the reliability alpha for the American identity centrality and Muslim identity centrality subscales were .87 and .85, respectively.

The 18-item Sense of Belonging Instrument – Psychological (SOBI-P) (Haggerty & Patusky, 1995) was administered to assess a general sense of belonging. Participants were asked to rank each item on a four-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree). Sample items included: “I generally feel that people accept me” and “I don’t feel that there is any place where I really fit in this world.” Items that indicate less belonging were reverse-coded so that higher scores on this scale indicated a greater sense of belonging. Total scores were calculated by summing the values of each item and the range was 18–72. The reliability alpha was .92.

To assess for American society belonging and Muslim community belonging, participants were asked to indicate their responses on a Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree) for the following items: “I feel belonging to the general American society” and “I feel belonging to my religious community.” Each item assessed for American society belonging and Muslim community belonging, respectively.

The World Health Organization’s Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST; Humeniuk et al., 2010) was used to assess participants’ substance use status and patterns. The first question prompted participants to identify whether they have ever used the following substances for non-medical use: tobacco products, alcoholic beverages, cannabis, cocaine, amphetamines, inhalants, sedatives or sleeping pills, hallucinogens, and opioids. If participants indicated using any of the prompted substances, the second question asked participants to indicate how often they used that substance within the last three months. Participants reporting use in the past three months were then asked three more questions per substance, which included how often they “felt a strong desire or urge to use,” how often their use “led to health, social, legal, or financial problems,” and how often they “failed to do what is normally expected of you because of your use.” These four items were scored on a 5-point Likert scale (for example, never, once or twice, monthly, weekly, and daily or almost daily). Finally, all participants who reported any lifetime use were asked three additional questions regarding if a “friend or relatives or anyone else ever expressed concern about (their) use” of each substance, if the participant had “ever tried and failed to control, cut down, or stop using” each substance reported, and if the participant had, through non-medical use, “ever used any drug by injection.” All three items were scored on a 3-point Likert scale (for example, never; yes, but not in the past 3 months; and yes, in the past 3 months) and were scored to determine low, moderate, or high-risk levels of use as suggested by the ASSIST scoring guide. The ASSIST measure was used in two ways. First, the frequency of lifetime and past 3-month use, as well as the risk level of use, are included in the frequency table below. Second, lifetime substance usage or whether someone ever used any substance in their lifetime, was an individual score used in the analysis. Namely, participants who indicated any substance use in their lifetimes were scored as 1 on this item, whereas participants who indicated no substance use in their lifetimes were scored as -1.

The frequency of substance use is displayed in Table 1, showing lifetime use and past 3-month use, as well as the risk level of each user. Tobacco use was most reported, with 32.8% (n = 60) of the sample reporting at least one use in the past and 10.4% being at moderate risk levels. Next, 26.8% (n = 49) of the sample reported using cannabis at least once in the past, with 8.7% of the sample at a moderate risk level; 15.8% (n = 29) of the sample reported drinking alcohol at least once, with 1.1% of the sample at a moderate risk level; and 4.4% (n = 8) and 3.3% (n = 6) of the sample reported using sedatives and hallucinogens, respectively, at least once in their lifetime.

Table 2 displays the listwise Pearson’s correlation matrix. Those who endorsed a stronger American identity centrality were also more likely to report a stronger sense of belonging to American society (r = .39, p < 0.001), and participants who endorsed a stronger sense of Muslim identity centrality were also more likely to report a stronger sense of belonging to the Muslim community (r = .65, p < 0.001). Additionally, participants who reported a greater general sense of belonging were more likely to identify as male (r = .25, p < 0.001), be older (r = .18, p < 0.05), report higher socioeconomic status (r = .23, p < 0.01), stronger religious identity centrality (r = .16, p < 0.05), a stronger sense of belonging to American society (r = .35, p < 0.001), and a stronger sense of belonging to the Muslim community (r = .39, p < 0.001). Lastly, lifetime substance use, measured as whether someone has ever used substances in their lifetime or not, was negatively and significantly linked to a Muslim identity centrality (r = –.20, p < 0.01), and a sense of belonging to the Muslim community (r = –.27, p < 0.001). In other words, participants who reported using substance use at least once in their lifetime were less likely to report a stronger Muslim identity centrality and less likely to report a stronger sense of belonging to the Muslim community than those who never used any substances in their lifetime.

Binary logistic regression was used to predict substance use for Muslim American emerging adults. Due to evidence that points to gender (Bradby, 2007), age (Arnet, 2005), and socioeconomic status (Patrick et al., 2012) as predictors of substance use, these variables were included in the model. Additionally, cultural factors, namely American and Muslim identity centrality, were included in the model to control for their overlapping variance with the variables of interest found in the correlation matrix. Various forms of belonging were also included in the model, including a general sense of belonging, American society belonging, and Muslim community belonging. Assumptions for linearity (for example, a linear relationship between unstandardized predicted value and studentized residuals) and multicollinearity (for example, tolerance greater than .1 and VIF less than 10) were met, and there were no outliers found. The full model showed a statistically significant better fit than the constant-only model (χ²(8) = 20.49, p <.01). The model explained 14.6% (Nagelkerke R²) of the variance and correctly classified 62.9% of cases. See Table 3 for regression coefficients, Wald statistics, odds ratios, and 95% confidence intervals. Of all the predictors, Muslim community belonging was the only statistically significant predictor of lifetime substance use (OR = .73, 95% CI [.54, .99]). Namely, those who reported belonging to the Muslim community were less likely to report ever using substances in their lifetime.

There are several limitations to note in the present study. Due to the staunch social and religious disapproval of substance use, social desirability bias is likely in our sample, with participants underreporting their substance use (Jozaghi et al., 2016). In fact, response bias, which is common in surveys of Muslim Americans’ substance use behaviors, was also evident in our study (Arfken et al., 2013). Some potential participants opted out of responding to substance use questions in the survey and informed the researchers that they skipped the questions inquiring about their substance use patterns. Also, other potential participants refused to take the survey altogether because they felt documenting their substance use patterns could incriminate them, even after confidentiality was thoroughly explained.

In addition to the concerns related to collecting sensitive information, there were two sampling concerns to note. First, there is no comprehensive representative membership list of the Muslim American population in the United States or even at a state level; therefore, we used a snowball and convenience sampling method to collect data, which proved to be an effective sampling method for minorities despite its non-random nature (Hughes et al., 1995). Second, there is an overrepresentation of Muslim Americans of Middle Eastern, North African, and South Asian backgrounds and an underrepresentation of Black Muslim Americans. This breakdown makes the results of our study difficult to generalize to the Muslim American community, especially because ethnic and cultural differences within the Muslim American community can also be strong influences on behavior (Atterbeen et al., 2019).

Future research must take careful methodological considerations when collecting substance use patterns for young adult Muslim Americans. Previous qualitative research indicates that methodological issues in collecting data on substance use for Muslim Americans can be complex due to the accuracy in reporting stigmatized behavior (Dotinga, 2005). Namely, researchers must be careful about confidentiality and messaging around collecting this information because the identity of the researcher can impact social desirability in reports. This consideration must also extend to mental health resources for substance abuse among young Muslims. Arfken and Ahmed (2016) recommend utilizing online sources of prevention and intervention to reduce the stigma around substance abuse for Muslims. Demographic and cultural factors can be examined as moderators in the relationship between substance use and various health outcomes for Muslim Americans, such as gender. Additionally, ethnic and cultural factors, such as stigma and awareness of drug use, can be examined as risk or protective factors of substance use.

Previous research also showed that Muslim Americans are less likely to seek help for substance use issues than their non-Muslim peers (Abu-Ras et al., 2010). The low help-seeking intentions could be due to the lack of Muslim-specific interventions to address substance abuse concerns in a culturally sensitive and responsive manner. A report from the Substance Abuse and Mental Health Administration Services (2020) recommends increased availability of faith-based resources to enhance the protective effects of community belonging. To protect against the negative effects of substance use as a risk behavior for young Muslim Americans, culturally responsive interventions should focus on fostering a sense of belonging among emerging adults in the Muslim American community. Understanding how to regain that sense of belonging is especially key in addiction recovery, where the majority of effective therapy relies on building a community around the user that liberates them from the sense of shame to which they may have been exposed. Given the competing influences in social and cultural aspects of young adult Muslim life, future research should continue to examine the cultural context of Muslim Americans as it relates to the impact of belonging and community on other risk behaviors, such as suicidal ideation and help-seeking intentions.