Abstract
Purpose
This longitudinal study aimed to identify variation by race in the associations between religious involvement and initiation of alcohol, cigarette, and marijuana use, including distinctions by substance or type of religious involvement, in Black and White adolescent girls.
Methods
Data were drawn from interviews conducted at ages 11 through 17 with 2172 Pittsburgh Girls Study participants (56.8% Black; 43.2% White). Two indicators of public religious involvement, religious service attendance and participation in other religious activities, and two indicators of private religious involvement, prayer, and importance of religion were queried. A series of Cox proportional hazards regression analyses were conducted to identify independent effects of religious involvement indicators on initiation of each substance.
Results
Prior to adjusting for socioenvironmental and psychosocial factors (e.g., parental monitoring), importance of religion predicted initiation of alcohol use across race and cigarette and marijuana use in White but not Black girls. Participation in other religious activities also predicted marijuana use initiation only in White girls. In adjusted models, importance of religion remained significant for cigarette use initiation in White girls (hazard ratio [HR] = 0.68, 95% confidence intervals [CI]: 0.53–0.88) and participation in other religious activities remained significant for marijuana use initiation in White girls (HR = 0.63, CI: 0.47–0.83).
Conclusions
The protective effects of religious involvement against cigarette and marijuana use initiation are more robust for White than Black adolescent girls and overall relatively weak for alcohol use initiation. Furthermore, importance placed on religion may be a better indicator than religious service attendance of risk for adolescent substance use initiation.
Keywords: Religious involvement, Alcohol, Cigarettes, Marijuana, Adolescent girls, Black/African American
Introduction
Religious involvement is a well-documented protective factor against adolescent substance use [1, 2]. High religiosity and frequent engagement in religious activities have been linked to reduced risk for any alcohol consumption, cigarette smoking, and marijuana use [3–6], lower frequency of alcohol, cigarette, and marijuana use [3, 7–9], and later onset of alcohol [10] and marijuana use [11]. Potential distinctions between Black and White youth in the magnitude of these associations merit attention because of the known distinctions by race in religious involvement as well as the prevalence of alcohol, cigarette, and marijuana use. Black adolescents tend to rate religion as more important and engage in religious activities more frequently than their White counterparts [6, 12]. Alcohol use and cigarette smoking are less common among Black than White adolescents [13], but the prevalence of marijuana use is somewhat higher in Black adolescents [13], suggesting that the relationship between religious involvement and adolescent substance use may vary by race, substance, or both. Identifying these variations can inform prevention efforts, revealing for whom and against which outcomes religious involvement may be protective.
Public versus private religious involvement
Type of religious involvement—commonly categorized in the literature as public or private—is a key component to consider in assessing the potential protective effects of religious involvement against adolescent substance use. The mechanisms linking religious involvement with health behaviors are complex [14], but public and private religious involvement are broadly conceptualized as reflecting different (although commonly overlapping) pathways to adolescent behavior. Public religious involvement, typically operationalized as religious service attendance but also encompassing other structured religious activities, such as youth groups or Bible study classes, is believed to impact adolescent behavior through community ties, e.g., community-level monitoring of youth activities. Private religiosity, typically measured by frequency of prayer or global rating of importance of religion [7], is believed to impact adolescent behavior through development of a moral code of conduct. Although less commonly included than religious service attendance in substance use studies, private religious involvement has been linked to lower risk for adolescent alcohol, cigarette, and marijuana use [6, 15], though not consistently [16] and with some variation by substance [5]. Findings from the few studies that have examined both types of religious involvement in combination suggest that they may contribute independently to reduced risk for adolescent substance use [4, 6, 15].
The link between religious involvement and substance use in Black versus White youth
Among the few investigations of potential differences between Black and White youth in the association between alcohol use and religious involvement (primarily religious service attendance), there is some support for protective effects being specific to White youth. Wallace et al.’s study [6] based on Monitoring the Future data revealed protective effects of religious service attendance against past 30-day alcohol use in White but not Black adolescents. Similarly, Agrawal et al. [10] found lower lifetime prevalence of alcohol use and lower likelihood of early alcohol use initiation in White but not Black young women reporting weekly or more frequent service attendance during childhood. However, in Steinman et al.’s study [8] of over 30,000 Ohio high school students, weekly engagement in religious activities (service attendance or other activities) was associated with reduced frequency of past year alcohol use irrespective of race.
Research on race differences in the protective effects of religious involvement against cigarette or marijuana use is even scarcer and also somewhat mixed. In the studies referenced above, Wallace et al. found protective effects of religious service attendance against past 30 day cigarette smoking in White adolescents only, whereas Steinman et al. found protective effects against past year cigarette use across race. Both studies also reported an association between religious service attendance (or other religious activities) and reduced frequency of past year marijuana use among both Black and White adolescents. Similarly, a more recent study based on National Survey on Drug Use and Health data found a lower prevalence of lifetime adolescent marijuana use among adolescents who scored high on a religiosity index derived from both public and private religious involvement indicators [17].
Notably, the timing of initiation in relation to religious involvement, which has only rarely been examined in the larger literature [3, 11], was examined in only one of these studies assessing for race differences [10]. Thus, whether religious involvement may delay onset of substance use in adolescents and specifically whether these effects differ for Black versus White youth remains largely unknown.
Study aims
The current study aimed to identify variation by race in the associations between religious involvement and initiation of alcohol, cigarette, and marijuana use, including distinctions by substance or type of religious involvement, in Black and White adolescent girls. We examined these links in the context of substance use related risk factors that are also associated with religious involvement: socioeconomic and neighborhood factors [18–22], low parental monitoring [23–25], and conduct problems [26–28], by including them as covariates in the models. The present investigation focused on adolescent girls, given the evidence for greater religious involvement [29] and possible stronger links between religious involvement and substance use in female versus male youth [9, 30]. Based on the support (albeit mixed) for protective effects of religious service attendance being specific to White youth, we hypothesized that high public religious involvement—operationalized as religious service attendance and participation in other religious activities—would be more strongly associated with delayed substance use initiation in White than Black girls. Given the absence of support for race differences in the associations between private religious involvement and substance use initiation, we anticipated that private religious involvement—operationalized as prayer and selfrated importance of religion—would be associated with later substance use initiation across race. In addition, we expected to observe independent effects of public and private religious involvement on timing of substance use initiation—evidenced by statistically significant hazard ratios for public and private religious involvement variables when included in the same model—but did not hypothesize substance-specific associations.
Methods
Participants
The Pittsburgh Girls Study (N = 2450) is a longitudinal study of behavioral and emotional development in an urban population-based sample of girls enrolled in middle childhood. Participants were recruited in 1999–2000, following a stratified random sampling of 103,238 households in Pittsburgh: all households in the 23 lowest income neighborhoods and 50% of households in the additional 66 neighborhoods, to identify girls ages 5–8 years in assessment wave 1. Families in which the mother was unable to speak English or the girl was severely developmentally delayed or hearing impaired with no sign language skills were excluded. Baseline interviews were completed by 85.2% of eligible families. Girls and their primary caregivers (94% mothers) were assessed annually. Sample retention was high: 88.5% on average over the years that data used in the current analyses were collected (2003–2010), when girls were 11–17 years of age. (See Hipwell et al. [31] and Keenan et al. [32] for details of sample ascertainment.) Given our interest in differences between Black and White girls, we excluded the small subsample of girls (n = 144) identified as a race other than ‘Black/ African American’ or ‘White’ by their primary caregiver in the demographic section of the wave 1 interview. An additional 134 participants who were missing alcohol, cigarette, and marijuana use data across waves were also excluded, resulting in a final analytic sample of 2172 girls (56.8% Black, 43.2% White). A larger proportion of White than Black participants were excluded (7.1% vs. 4.8%; χ2(1) = 5.75; p = 0.017), reflecting the overall attrition pattern.
Procedures
Highly trained research staff conducted face-to-face interviews in participants’ homes, separately for the girl and her primary caregiver. Written informed consent from the primary caregiver and verbal assent from the child were obtained. The protocol for maintaining confidentiality was explained to all participants. The girls were reminded throughout the interview that their responses would not be shared with their caregivers. Families were compensated for their participation. The University of Pittsburgh’s Human Research Protection Office approved the protocol.
Assessment
Substance use
Alcohol, cigarette, and marijuana use were assessed starting at age 11, by querying past year frequency and usual quantity of use. For each substance, age at initiation was calculated as the age at first endorsement of any past year use.
Religious involvement
Two indicators of public religious involvement and two indicators of private religious involvement were assessed annually starting at age 13: (1) Frequency of religious service attendance: ‘In the past year, how often did you attend religious services?’ (4-level: ‘Never’ to ‘More than once a week’); (2) Participation in religious activities (other than service attendance): ‘About how often do you spend time in organized religious activities?’ (4-level with six original response options, from ‘Never’ to ‘Just about every day’ with three highest frequency options collapsed into ‘Once a week or more’ due to very low endorsement of the two highest frequency response options). (3) Importance of religion: ‘How important is religion to you?’ (3-level: ‘Not important’, ‘Somewhat important’, ‘Very important’); and (4) Frequency of prayer: ‘In the past year, how often did you pray?’ (4-level: ‘Never’ to ‘More than once a week’). For all indicators, higher values indicate higher frequency or degree of importance. Year by year endorsements of each religious involvement indicator are reported by race in Supplemental Table 1. Correlations among religious involvement indicators are shown by year, separately by race, in Supplemental Table 2.
Covariates (assessed at ages 11–17)
Socioeconomic status (primary caregiver report) was operationalized using binary variables representing household receipt of public assistance (0 = no, 1 = yes), single parentheaded household (0 = no, 1 = yes), and low primary caregiver educational attainment (<12 vs. ≥ 12 years).
Neighborhood factors covered three domains. (1) Low neighborhood safety (primary caregiver report) was assessed with Your Neighborhood [33], a 17-item measure of neighborhood problems (e.g., ‘delinquent gangs’) using a 3-point Likert scale (‘Not a problem’ to ‘Big problem’; possible range = 17–51, with higher scores indicating lower safety). Alphas ranged from 0.95 to 0.96 across waves. (2) Community cohesion (primary caregiver report) was assessed with 10 items from the Community Survey [34], a 13-item measure (e.g., ‘I regularly stop and talk with people in my neighborhood’) using a 5-point Likert scale (‘Strongly agree’ to ‘Strongly disagree’), with a possible range of 10–50, reverse coded so that higher scores indicate greater cohesion. Alphas ranged from 0.91 to 0.92 across waves. (3) Neighborhood physical disorder (study interviewer report) was indexed with five dichotomously coded items (e.g., presence of graffiti), from the 25-item Interviewer Impressions of the Neighborhood [35]. Possible scores ranged from 0 to 5, with higher scores indicating greater physical disorder. Alphas ranged from 0.69 to 0.73 across waves.
Low parental monitoring (child report) was assessed with four items (e.g., ‘When you are out, do your parent(s) know what time you will be home?’) from the Supervision and Involvement Scale [33], rated on a 3-point scale (‘Almost always’ to ‘Almost never’; possible range = 4–12, with higher scores indicating lower monitoring). Alphas ranged from 0.61 to 0.74 across waves.
Conduct problems (child report) were assessed using the Adolescent Symptom Inventory-4 [36], which queries frequency of 15 past year conduct disorder symptoms, using a 4-point scale (‘Never’ to ‘Very often’) to generate a severity score (possible range = 0–45, with higher scores indicating greater severity). Alphas ranged from 0.64 to 0.75 across waves.
Age 11 socioeconomic status indicators, neighborhood factors, and scores on parental monitoring and conduct problem assessments are reported by race in Tables 1 and 2.
Table 1.
Sample characteristics by race: socioeconomic status indicators, neighborhood factors, low parental monitoring, and conduct problems at age 11
| Black girls (n = 1235) | White girls (n = 937) | |
|---|---|---|
| Household receipt of public assistance | 49.6% | 15.3% |
| Single parent headed household | 59.4% | 21.7% |
| Primary caregiver education < 12 years | 17.4% | 11.5% |
| Low neighborhood safety: M (SD) | 25.95 (8.71) | 20.69 (5.97) |
| Community cohesion: M (SD) | 31.13 (8.48) | 37.72 (8.08) |
| Neighborhood physical disorder: M (SD) | 1.43 (1.42) | 0.49 (0.90) |
| Low parental monitoring: M (SD) | 4.83 (1.27) | 4.43 (0.91) |
| Conduct problems: M (SD) | 1.11 (1.68) | 0.46 (0.95) |
Socioeconomic status indicators and neighborhood factors are primary caregiver report reports; parental monitoring and conduct problems are child reports
M mean, SD standard deviation
Differences across race are statistically significant at p < 0.05 for all variables
Table 2.
Lifetime prevalence and age at initiation of alcohol, cigarette, and marijuana use by race
| Any use (%) |
Age at first use: M (SD) |
|||||
|---|---|---|---|---|---|---|
| Black girls | White girls | χ2(1) | Black girls | White girls | t(df) | |
| Alcohola,b | 52.8 | 67.7 | 48.62 | 14.08 (2.00) | 13.70 (1.98) | t(1282) = 3.42 |
| Cigarettesa | 26.2 | 34.7 | 18.16 | 14.53 (2.02) | 14.70 (1.75) | t(633.85) = 1.15 |
| Marijuanaa,b | 41.8 | 31.2 | 25.53 | 14.99 (1.58) | 15.33 (1.36) | t(682.73) = 3.24 |
M mean, SD standard deviation
Difference in prevalence significant at p < 0.05
Difference in age at first use significant at p < 0.05
Analytic approach
Cox proportional hazards (PH) regression analyses were conducted in stages to predict alcohol, cigarette, and marijuana use initiation as a function of religious involvement, in separate series of analyses for each substance. This survival analysis approach accounts for the possibility that participants may not have passed through the period of risk and thus is well suited for an adolescent sample. All predictors and covariates were entered as time-varying variables. Religious involvement questions were not included at ages 11 and 12. Rather than limiting the analyses to initiation from ages 13 to 17, thus losing early initiators, we assigned the age 11 and 12 religious involvement variables the same value as the age 13 variables, given their relative stability over time. (Within religious indicator correlations at 1-year and 2-year lags, respectively, ranged from 0.54 to 0.68 and 0.50 to 0.57 for religious service attendance, 0.67 to 0.82 and 0.64 to 0.74 for importance of religion, 0.63 to 0.76 and 0.61 to 0.68 for prayer, and 0.57 to 0.70 and 0.50 to 0.62 for participation in religious activities.) Analyses were conducted in Stata [37]. Violations of the PH assumption that risk remains constant over time were resolved by splitting the period of risk and estimating hazard ratios for each risk period.
The preliminary stage of analysis consisted of two steps. In the first step, we established whether each of the religious involvement variables was associated with initiation by modeling them individually, along with race. In the second step, all significant religious involvement variables were entered, along with race, into a single model. To derive the most parsimonious model, only religious involvement variables that remained significant in step two were included in the primary analyses. For the primary analyses, results of unadjusted models (Model 1), which included only race and religious involvement variables, as well as adjusted models (Model 2), which also included socioeconomic status indicators, neighborhood factors, low parental monitoring, and conduct problems, are presented in Tables 3 and 4. Comparing results of Models 1 and 2 demonstrates the effects of adding covariates on the hazard ratio estimates for religious involvement and race.
Table 3.
Results of Cox proportional hazards regression analyses predicting initiation of alcohol use
| Model 1 HR (95% CI) |
Model 2 HR (95% CI) |
|
|---|---|---|
| Importance of religion | 0.84 (0.78–0.91) | 0.94 (0.87–1.03) |
| Race: Black compared to White | 0.65 (0.58–0.73) | 0.57 (0.49–0.66) |
| Socioeconomic status indicators | ||
| Household receipt of public assistance | – | 0.88 (0.77–1.02) |
| Single parent headed household | – | 0.94 (0.82–1.07) |
| Primary caregiver education < 12 years | – | 0.83 (0.69–0.99) |
| Neighborhood factors | – | |
| Low neighborhood safety | – | 1.01 (1.00–1.01) |
| Community cohesion | – | 1.01 (1.00–1.01) |
| Neighborhood physical disorder | – | 0.99 (0.94–1.04) |
| Low parental monitoring | – | 1.08 (1.03–1.13) |
| Conduct problems | – | 1.23 (1.21–1.26) |
Model 1: importance of religion and race; Model 2: addition of socioeconomic status indicators, neighborhood factors, low parental monitoring, and conduct problems
Bold indicates statistically significant at p < 0.05
HR hazard ratio, CI confidence interval
Table 4.
Results of Cox proportional hazards regression analyses predicting initiation of cigarette use
| Model 1 HR (95% CI) |
Model 2 HR (95% CI) |
|
|---|---|---|
| Importance of religiona | ||
| Black girls | 0.79 (0.56–1.12) | 1.26 (0.87–1.85) |
| White girls | 0.46 (0.36–0.58) | 0.68 (0.53–0.88) |
| Race: Black compared to Whiteb | ||
| Initiation of cigarette smoking ≤ age 14 | 0.48 (0.33–0.71) | 0.21 (0.13–0.34) |
| Initiation of cigarette smoking ≥ age 15 | 0.28 (0.18–0.44) | |
| Socioeconomic status indicators | ||
| Household receipt of public assistance | – | 1.18 (0.97–1.43) |
| Single parent headed household | – | 1.07 (0.89–1.29) |
| Primary caregiver education < 12 years | – | 1.06 (0.83–1.34) |
| Neighborhood factors | ||
| Low neighborhood safety | – | 1.01 (1.00–1.01) |
| Community cohesion | – | 1.01 (1.00–1.01) |
| Neighborhood physical disorderb | – | |
| Initiation of cigarette smoking ≤ age 14 | 1.12 (1.01–1.25) | |
| Initiation of cigarette smoking ≥ age 15 | 0.95 (0.86–1.06) | |
| Low parental monitoring | ––- | 1.09 (1.02–1.15) |
| Conduct problems | ––- | 1.30 (1.27–1.34) |
Model 1: importance of religion and race; Model 2: addition of socioeconomic status indicators, neighborhood factors, low parental monitoring, and conduct problems
Bold indicates statistically significant at p < 0.05
HR hazard ratio, CI confidence interval
Separate estimates generated for Black and White girls, given significant interaction with race
Risk period split to adjust for proportional hazards violations in Model 2
Results
Alcohol, cigarette, and marijuana use and religious involvement by race
As seen in Table 2, lifetime prevalence of alcohol and cigarette use was significantly lower in Black girls than White girls (52.8% vs. 67.7% and 26.2% vs. 34.7%, respectively), whereas marijuana use was higher (41.8% vs. 31.2%). Black girls reported an older mean age at first alcohol use (14.08 years, SD = 2.00 vs. 13.70 years, SD = 1.98) and a younger mean age at first marijuana use (14.99 years, SD = 1.58 vs. 15.33 years, SD = 1.36) than White girls. No difference by race in age at first use of cigarettes was observed. Logistic regression analyses using race as a predictor of maximum values of religious involvement indicators (across years assessed) revealed greater frequency of religious service attendance [odds ratio (OR) = 1.80, 95% confidence intervals (CI): 1.53–2.11] and prayer (OR = 2.21, CI: 1.86–2.63) and higher rating of importance of religion (OR = 2.43, CI: 2.02–2.92) in Black compared to White girls, but no difference by race in frequency of participation in other religious activities (OR = 1.11, CI: 0.94–1.30). Lifetime prevalence of alcohol, cigarette, and marijuana use by religious involvement at age 17 are reported by race in Supplemental Table 3.
Initiation of alcohol use as a function of self-reported importance of religion
Cox PH regression analyses conducted separately for each religious involvement variable (step 1) revealed a significant association exclusively with importance of religion. Thus, only importance of religion was included in the primary modeling stage. As seen in Table 3, in Model 1, importance of religion was associated with delayed initiation of alcohol use [hazard ratio (HR) = 0.84, CI: 0.78–0.91]. The increase in the HR for importance of religion from Models 1 to 2 was statistically significant (the point estimate was outside the CIs in Model 1) and the HR in Model 2 was non-significant (HR = 0.94, CI: 0.87–1.03). Low primary caregiver education was associated with delayed alcohol use initiation (HR = 0.83, CI: 0.69–0.99), whereas low parental monitoring (HR = 1.08, CI: 1.03–1.13) and conduct problems (HR = 1.23, CI: 1.21–1.26) were associated with earlier initiation.
Initiation of cigarette use as a function of self-reported importance of religion
Each of the four religious involvement variables predicted cigarette use initiation in separate Cox PH regression analyses (step 1), but importance of religion was the only one retained from the model containing all four (step 2). Interactions between race and importance of religion were significant in both Models 1 and 2. As shown in Table 5, separate HR estimates were therefore derived for Black and White participants. In Model 1, importance of religion was associated with delayed cigarette use initiation exclusively in White girls (HR = 0.46, CI: 0.33–0.71). In Model 2, the increase in HR for White girls from 0.46 (CI: 0.36–0.58) to 0.68 (CI: 0.53–0.88) for importance of religion was statistically significant. Separate estimates for the risk periods ≤ age 14 and ≥ age 15 were derived for race to adjust for PH violations in Model 2. The same split was applied to adjust for PH violations for neighborhood physical disorder, with elevated risk observed for the ≤ age 14 risk period (HR = 1.12, CI: 1.01–1.25). Low parental monitoring (HR = 1.09, CI: 1.02–1.15) and conduct problems (HR = 1.30, CI: 1.27–1.34) were also associated with earlier cigarette use initiation.
Table 5.
Results of Cox proportional hazards regression analyses predicting initiation of marijuana use
| Model 1 HR (95% CI) |
Model 2 HR (95% CI) |
|
|---|---|---|
| Importance of religiona | ||
| Black girlsb | ||
| Initiation of marijuana use ≤ age 15 | 0.80 (0.59–1.09) | 1.38 (1.00–1.91) |
| Initiation of marijuana use ≥ age 16 | 0.95 (0.69–1.32) | |
| White girls | 0.60 (0.46–0.79) | 0.84 (0.63–1.11) |
| Participation in religious activitiesa,c | ||
| Black girls | 0.92 (0.77–1.10) | 1.07 (0.88–1.29) |
| White girls | 0.51 (0.39–0.67) | 0.63 (0.47–0.83) |
| Race: Black compared to White | 0.87 (0.63–1.22) | 0.55 (0.38–0.81) |
| Socioeconomic status indicators | ||
| Household receipt of public assistance | – | 1.05 (0.89–1.24) |
| Single parent headed household | – | 1.07 (0.91–1.26) |
| Primary caregiver education < 12 years | – | 0.92 (0.74–1.13) |
| Neighborhood factors | ||
| Low neighborhood safety | – | 1.00 (0.99–1.01) |
| Community cohesion | – | 1.01 (0.99–1.01) |
| Neighborhood physical disorder | – | 1.04 (0.98–1.11) |
| Low parental monitoring | – | 1.08 (1.02–1.13) |
| Conduct problems | – | 1.28 (1.25–1.30) |
Model 1: importance of religion, participation in religious activities, and race; Model 2: addition of socioeconomic status indicators, neighborhood factors, low parental monitoring, and conduct problems
Bold indicates statistically significant at p < 0.05
HR hazard ratio, CI confidence interval
Separate estimates generated for Black and White girls, given significant interaction with race
Risk period split to adjust for proportional hazards violations in Model 1
Religious activities other than service attendance
Initiation of marijuana use as a function of self-reported importance of religion and participation in other religious activities
Each of the four religious involvement variables predicted marijuana use initiation in separate Cox PH regression analyses (step 1). Importance of religion and participation in other religious activities were retained from the model containing all four (step 2). Interactions with race were significant, so separate HR estimates were derived for Black and White participants for both importance of religion and participation in other religious activities. In Model 1, in White but not Black girls, importance of religion and participation in other religious activities were associated with later marijuana use initiation. The HRs for White girls were 0.60 (CI: 0.46–0.79) for importance of religion and 0.51 (CI: 0.39–0.67) for participation in other religious activities. The HRs for importance of religion increased significantly in both racial groups from Models 1 to 2, becoming non-significant for White girls (and remaining non-significant for Black girls). The HR specific to White girls for participation in religious activities did not change significantly and remained statistically significant (HR = 0.63, CI: 0.47–0.83). Low parental monitoring (HR = 1.08, CI: 1.02–1.13) and conduct problems (HR = 1.28, CI: 1.25–1.30) were also associated with earlier initiation of marijuana use in Model 2.
Discussion
The present study advances our understanding of the link between religious involvement and adolescent substance use initiation by identifying differences in the association across race, substance, and type of involvement in Black and White adolescent girls. Findings support self-rated importance of religion as a prominent protective factor against early substance use initiation, with distinctions by race and substance in the robustness of those associations after considering correlated socioenvironmental and psychosocial factors. By contrast, study findings do not support independent effects of religious service attendance on substance use initiation in Black or White adolescents when considered in combination with self-rated importance of religion.
Association of public religious involvement with substance use initiation
Consistent with the extant literature, in the current study, Black girls reported higher frequency of religious service attendance and prayer and greater importance of religion [6, 12] as well as lower prevalence of alcohol and cigarette use but higher prevalence of marijuana use [13] than White girls. However, our hypothesis that public religious involvement, i.e., religious service attendance and/or participation in other religious activities, would be more strongly associated with delayed substance use initiation in White compared to Black girls was minimally supported. Participation in religious activities other than religious services was associated with later onset of marijuana use only in White girls, but it was not associated with alcohol or cigarette use initiation in either White or Black girls. Furthermore, frequency of religious service attendance did not predict initiation of any substance use in either group. Among the few studies assessing race differences in the link between public religious involvement and substance use, some have found associations specific to White youth for alcohol or cigarettes [6, 10], but those examining marijuana use have found protective effects across race [6, 8].
In the absence of similar findings in the existing literature, we are left to speculate on why protective effects of public religious involvement against substance use initiation were specific to White girls, marijuana, and participation in religious activities other than service attendance. As participation in other religious activities was the one form of religious involvement that was not more highly endorsed by Black than White girls, we cannot make the argument that protective effects were not observable in Black girls because this form of public religious involvement was nearly ubiquitous. One possibility is that participation in youth religious activities, which sometimes integrate substance use prevention efforts, may impact marijuana use in particular, given the stigma associated with illicit drug use among highly religious individuals [38], and White youth in particular, as marijuana use is somewhat less normative for White than Black youth.
Association of private religious involvement with substance use initiation
Our second hypothesis, that private religious involvement, i.e., high importance of religion and/or frequency of prayer, would protect against early substance use initiation across race, was also partially supported. Although frequency of prayer did not emerge as an independent predictor in the context of the other religious involvement indicators, consistent with prior work in this area [4, 6, 9, 11, 15], importance of religion, a more global indicator of religious devotion, predicted initiation of alcohol, cigarette, and marijuana use in unadjusted models. However, counter to findings from the one known study to explore race differences in the link between private religious involvement and substance use [6], associations with cigarette and marijuana use were specific to White girls. In contrast with participation in other religious activities, which did not vary by race and was endorsed by a minority of participants across ages, as seen in Supplemental Table 1, approximately 90% of Black girls rated religion as at least somewhat important, 50% as very important, with greater variability across ages and lower overall rates among White girls (endorsement of at least somewhat important dropped from 86 to 70% and very important from 46 to 29% from ages 13 to 17). Thus, importance of religion may not differentiate levels of risk for substance use initiation as well among Black compared to White girls.
Also of note, hazard ratios were significantly attenuated in all three models after adjusting for correlated socioenvironmental and psychosocial factors and importance of religion remained significant exclusively in the cigarette model. Estimates of additional risk conferred by low parental monitoring and conduct problems were comparable across substances (hazard ratios ranged from 1.08 to 1.09 for low parental monitoring and 1.23 to 1.30 for conduct problems), but findings for socioenvironmental factors were more limited and substance-specific. Consistent with prior studies, low parental education was associated with later alcohol use initiation [39] and neighborhood physical disorder was associated with earlier initiation of cigarette smoking [40].
Independence of protective effects of public and private religious involvement
Support for our hypothesis that public and private religious involvement would be independently associated with delayed initiation of substance use was limited to results of the unadjusted model for marijuana use. In that model, both participation in other religious activities (a form of public religious involvement) and importance of religion (a form of private religious involvement) predicted later initiation in White girls. Although the specificity of effects to White girls is inconsistent with the one known study of the joint effects of public and private religiosity to examine differences between Black and White youth [6], our findings are in keeping more broadly with prior studies of adolescent marijuana use that investigated both public and private religious involvement [4, 11, 15, 41]. The absence of evidence for independent protective effects for alcohol and cigarette use initiation was somewhat surprising, as all the same studies that found independent effects for marijuana use also found them for cigarette use and most found them for alcohol use [4, 6, 11]. One plausible source of variation is the modeling of alcohol and cigarette smoking outcomes as onset of use in our study, compared to ‘current’ or past 30-day use in most prior studies [6, 15, 39], given that influences on initiation may differ from those on current patterns of use. Among this group of studies, two assessed past year marijuana use [6, 15] and one assessed lifetime marijuana use [41], which are more similar time frames to ours than the ‘current’ and past 30-day time frames used for alcohol and cigarettes.
Limitations
Our findings should be interpreted with certain limitations in mind. First, as religious involvement data were not available for ages 11 and 12, we estimated age 11 and 12 values as the same as those reported at age 13. The consecutive-year (and 2-year-lag) correlations suggest high continuity across ages, but accuracy is inherently lower for estimated than reported values. Second, religious involvement variables were derived from single items. Although a common method, certain aspects of religious involvement may be better captured with multi-item measures. Third, given our interest in predicting first experience with each substance, first use of alcohol could be ‘a sip or taste’, which could occur in the context of a religious ceremony. Fourth, the last wave of data collection included in the sample was conducted in 2010. The possibility that the association between religious involvement and adolescent substance use has changed since that time should be considered. Fifth, our aim was to identify links between religious involvement and timing of initiation of substance use; inferences cannot be made about associations with frequency or heaviness of use.
Conclusions and future directions
Our findings underscore the complexity of the link between religious involvement and adolescent substance use and suggest a number of potential directions for future research. Results indicate that protective effects of religious involvement against early cigarette and marijuana use initiation are more robust for White than Black adolescent girls and overall relatively weak for alcohol use initiation. In addition, they suggest that importance placed on religion is a stronger indicator than frequency of religious service attendance of risk for early substance use initiation, and thus, increasing frequency of religious service attendance would be unlikely to confer additional protection against early substance use initiation. The lesser protective effect of religious service attendance against substance use initiation likely reflects the fact that frequency of service attendance among adolescents is influenced by family and community norms in addition to personal beliefs, and for some adolescents, service attendance is not voluntary.
Key next steps in this line of research—in addition to assessing generalizability of findings to other subpopulations of adolescents (e.g., males, youth from other racial and ethnic backgrounds or non-urban environments)—include the investigation of heavy and problem substance use. Distinctions from our findings for initiation may emerge, given that substance use related problems typically develop in late adolescence and early adulthood, when many youth no longer reside with their family of origin, and thus, both public and private religious involvement are less heavily influenced by family practices and community norms. Testing for moderating effects of private by public religious involvement on substance use outcomes [5] could also provide insight into the conditions under which protective effects are most likely to manifest. Finally, conducting more comprehensive assessments of religious involvement aimed at identifying the core elements driving its protective effects against substance use initiation, such as maintaining family or cultural traditions or following a shared moral code of conduct, could inform prevention efforts.
Supplementary Material
Acknowledgements
Funding for this study was provided by the National Institute on Alcohol Abuse and Alcoholism (AA023549), the National Institute on Drug Abuse (DA012237), the National Institute on Mental Health (MH056630), FISA Foundation, and the Falk Fund. We thank all of the research participants and their families for the time which they dedicated to this study.
Footnotes
Compliance with ethical standards
Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00127–020-01830-y) contains supplementary material, which is available to authorized users.
References
- 1.Koenig HG, King D, Carson VB (2012) Handbook of religion and health, 2nd edn. Oxford University Press, New York [Google Scholar]
- 2.Weaver AJ, Flannelly KJ, Strock AL (2005) A review of research on the effects of religion on adolescent tobacco use published between 1990 and 2003. Adolescence 40:761–776 [PubMed] [Google Scholar]
- 3.Miller L, Davies M, Greenwald S (2000) Religiosity and substance use and abuse among adolescents in the National Comorbidity Survey. J Am Acad Child Adolesc Psychiatry 39:1190–1197 [DOI] [PubMed] [Google Scholar]
- 4.Nonnemaker JM, McNeely CA, Blum RW (2003) Public and private domains of religiosity and adolescent health risk behaviors: evidence from the National Longitudinal Study of Adolescent Health. Soc Sci Med 57:2049–2054 [DOI] [PubMed] [Google Scholar]
- 5.Pitel L, Geckova AM, Kolarcik P, Halama P, Reijneveld SA, van Dijk JP (2012) Gender differences in the relationship between religiosity and health-related behaviour among adolescents. J Epidemiol Commun Health 66:1122–1128 [DOI] [PubMed] [Google Scholar]
- 6.Wallace JM Jr, Delva J, O’Malley PM, Bachman JG, Schulenberg JE, Johnston LD, Stewart C (2007) Race/ethnicity, religiosity and adolescent alcohol, cigarette and marijuana use. Soc Work Public Health 23:193–213 [DOI] [PubMed] [Google Scholar]
- 7.Nonnemaker J, McNeely CA, Blum RW (2006) Public and private domains of religiosity and adolescent smoking transitions. Soc Sci Med 62:3084–3095 [DOI] [PubMed] [Google Scholar]
- 8.Steinman KJ, Ferketich AK, Sahr T (2008) The dose–response relationship of adolescent religious activity and substance use: variation across demographic groups. Health Educ Behav 35:22–43 [DOI] [PubMed] [Google Scholar]
- 9.Wills TA, Yaeger AM, Sandy JM (2003) Buffering effect of religiosity for adolescent substance use. Psychol Addict Behav 17:24–31 [DOI] [PubMed] [Google Scholar]
- 10.Agrawal A, Grant JD, Haber JR, Madden PAF, Heath AC, Bucholz KK, Sartor CE (2017) Differences between White and Black young women in the relationship between religious service attendance and alcohol involvement. Am J Addict 26:437–445 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hayatbakhsh R, Clavarino A, Williams GM, Najman JM (2014) Maternal and personal religious engagement as predictors of early onset and frequent substance use. Am J Addict 23:363–370 [DOI] [PubMed] [Google Scholar]
- 12.Brown TL, Parks GS, Zimmerman RS, Phillips CM (2001) The role of religion in predicting adolescent alcohol use and problem drinking. J Stud Alcohol 62:696–705 [DOI] [PubMed] [Google Scholar]
- 13.Johnston LD, Miech RA, O’Malley PM, Bachman JG, Schulenberg JE, Patrick ME (2018) Monitoring the Future National Survey results on drug use: 1975–2017: overview, key findings on adolescent drug use. Institute for Social Research, Ann Arbor [Google Scholar]
- 14.Hill PC, Pargament KI (2003) Advances in the conceptualization and measurement of religion and spirituality: implications for physical and mental health research. Am Psychol 58:64–74 [DOI] [PubMed] [Google Scholar]
- 15.Salas-Wright CP, Vaughn MG, Maynard BR, Clark TT, Snyder S (2017) Public or private religiosity: which is protective for adolescent substance use and by what pathways? Youth Soc 49:228–253 [Google Scholar]
- 16.Mason MJ, Schmidt C, Mennis J (2012) Dimensions of religiosity and access to religious social capital: correlates with substance use among urban adolescents. J Prim Prev 33:229–237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nguyen NN, Hewill CE (2016) The role of religiosity as a protective factor against marijuana use among African American, White, Asian, and Hispanic adolescents. J Subst Use 21:547–552 [Google Scholar]
- 18.Blum RW, Beuhring T, Shew ML, Bearinger LH, Sieving RE, Resnick MD (2000) The effects of race/ethnicity, income, and family structure on adolescent risk behaviors. Am J Public Health 90:1879–1884 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Franzini L, Ribble JC, Wingfield KA (2005) Religion, sociodemographic and personal characteristics, and self-reported health in Whites, Blacks, and Hispanics living in low-socioeconomic status neighborhoods. Ethnic Dis 15:469–484 [PubMed] [Google Scholar]
- 20.Pew Research Center (2017) In America, does more education equal less religion?
- 21.Shih RA, Parast L, Pedersen ER, Troxel WM, Tucker JS, Miles JNV, Kraus L, D’Amico EJ (2017) Individual, peer, and family factor modification of neighborhood-level effects on adolescent alcohol, cigarette, e-cigarette, and marijuana use. Drug Alcohol Depend 180:76–85 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tucker JS, Pollard MS, de la Haye K, Kennedy DP, Green HD Jr (2013) Neighborhood characteristics and the initiation of marijuana use and binge drinking. Drug Alcohol Depend 128:83–89 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Corwyn RF, Benda BB (2000) Religiosity and church attendance: the effects on use of “hard drugs” controlling for sociodemographic and theoretical factors. Int J Psychol Relig 10:241–258 [Google Scholar]
- 24.Steinberg L, Fletcher A, Darling N (1994) Parental monitoring and peer influences on adolescent substance use. Pediatrics 93:1060–1064 [PubMed] [Google Scholar]
- 25.Van Ryzin MJ, Fosco GM, Dishion TJ (2012) Family and peer predictors of substance use from early adolescence to early adulthood: an 11-year prospective analysis. Addict Behav 37:1314–1324 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bartkowski JP, Xu XH, Levin ML (2008) Religion and child development: evidence from the Early Childhood Longitudinal Study. Soc Sci Res 37:18–36 [Google Scholar]
- 27.Groenman AP, Janssen TWP, Oosterlaan J (2017) Childhood psychiatric disorders as risk factor for subsequent substance abuse: a meta-analysis. J Am Acad Child Adolesc Psychiatry 56:556–569 [DOI] [PubMed] [Google Scholar]
- 28.Krueger RF, Hicks BM, Patrick CJ, Carlson SR, Iacono WG, McGue M (2002) Etiologic connections among substance dependence, antisocial behavior, and personality: modeling the externalizing spectrum. J Abnorm Psychol 111:411–424 [PubMed] [Google Scholar]
- 29.Smith C, Denton ML, Faris R, Regnerus M (2002) Mapping American religious participation. J Sci Study Relig 41:597–612 [Google Scholar]
- 30.Oman RF, Vesely S, Aspy CB, McLeroy KR, Rodine S, Marshall L (2004) The potential protective effect of youth assets on adolescent alcohol and drug use. Am J Public Health 94:1425–1430 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hipwell AE, Loeber R, Stouthamer-Loeber M, Keenan K, White HR, Kroneman L (2002) Characteristics of girls with early onset disruptive and antisocial behaviour. Crim Behav Ment Health 12:99–118 [DOI] [PubMed] [Google Scholar]
- 32.Keenan K, Hipwell A, Chung T, Stepp S, Stouthamer-Loeber M, Loeber R, McTigue K (2010) The Pittsburgh Girls Study: overview and initial findings. J Clin Child Adolesc 39:506–521 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Loeber R, Farrington D, Stouthamer-Loeber M, Van Kammen W (1998) Antisocial behavior and mental health problems: explanatory factors in childhood and adolescence. Lawrence Erlbaum Associates, Mahwah [Google Scholar]
- 34.Gorman-Smith D, Tolan PH, Henry D (2000) A developmentalecological model of the relation of family functioning to patterns of delinquency. J Quant Criminol 16:169–198 [Google Scholar]
- 35.Wei E, Hipwell A, Pardini D, Beyers JM, Loeber R (2005) Block observations of neighbourhood physical disorder are associated with neighbourhood crime, firearm injuries and deaths, and teen births. J Epidemiol Community Health 59:904–908 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Gadow KD, Sprafkin J (1997) Adolescent Symptom Inventory-4. Checkmate Plus, Stony Brook [Google Scholar]
- 37.StataCorp (2007) Stata 92 edition. StataCorp, College Station [Google Scholar]
- 38.Palamar JJ, Kiang MV, Halkitis PN (2011) Development and psychometric evaluation of scales that assess stigma associated with illicit drug users. Subst Use Misuse 46:1457–1467 [DOI] [PubMed] [Google Scholar]
- 39.Kim YM, Neff JA (2010) Direct and indirect effects of parental influence upon adolescent alcohol use: a structural equation modeling analysis. J Child Adolesc Subst 19:244–260 [Google Scholar]
- 40.Diez Roux AV, Merkin SS, Hannan P, Jacobs DR, Kiefe CI (2003) Area characteristics, individual-level socioeconomic indicators, and smoking in young adults: the Coronary Artery Disease Risk Development in Young Adults Study. Am J Epidemiol 157:315–326 [DOI] [PubMed] [Google Scholar]
- 41.Sinha JW, Cnaan RA, Gelles RJ (2007) Adolescent risk behaviors and religion: findings from a national study. J Adolesc 30:231–249 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.