Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2010 Mar 11.
Published in final edited form as: Soc Work Res. 2008;32(3):147–157. doi: 10.1093/swr/32.3.147

Pathways to Drug and Sexual Risk Behaviors among Detained Adolescents

Dexter R Voisin 1, Torsten B Neilands 2, Laura F Salazar 3,5, Richard Crosby 4,5,6, Ralph J DiClemente 3,5,6,7,8
PMCID: PMC2836779  NIHMSID: NIHMS180906  PMID: 20228887

Abstract

Purpose

Among a sample of detained youth, to investigate pathways that link witnessing community violence, in the 12 months prior to being detained, to drug and sexual risk behaviors, in the two months preceding detainment.

Methods

Using A-CASI technology, data was collected from 559 detained adolescents on demographics, family factors, peer influences, religiosity, witnessing community violence, and drug and sexual risk behaviors.

Results

Controlling for demographics and family variables, findings indicated positive associations between witnessing community violence and drug and sexual risk behaviors. Witnessing community violence was directly linked to sexual risk behaviors, and indirectly associated with these risk behaviors, through gang membership and perceived risky peer norms. Additionally, witnessing community violence was indirectly linked to substance use through gang membership and perceived risky peer norms.

Conclusions

Interventions targeting change in peer affiliations and perceived norms may be an effective strategy for reducing risky drug and sexual behaviors among detained youth.

Keywords: Community violence, gangs, peer norms, drug use, HIV risks, detained youth

INTRODUCTION

In the United States considerable numbers of youth are involved in the juvenile justice system. Each year approximately 2.5 million youth are arrested (Snyder, 2003) and an additional 1.8 million cases referred to juvenile courts (Puzzanchera, 2003). Furthermore, daily averages of 109,000 youth (ages 18 and younger) are incarcerated (Snyder, 2003). Incarceration rates however, are not consistent across all adolescent populations. For example, the number of juvenile female detainees is increasing at a much faster rate than that of males (Sickmaund, Sladky, & Kang, 2003). In addition, African American and Hispanic youth while representing 20% of the adolescent population (U.S. Census, 2000) account for approximately 60% of youth detainees (Sickmaund, Sladky, & Kang, 2003).

Many detained (or previously detained) come from disadvantaged backgrounds. For instance, many of these youth live in impoverished communities which may place them at increased risk for exposure to negative influences and other dangers such as community violence (Canterbury et al., 1995; Chung & Steinberg, 2006). In this study, community violence exposure refers to witnessing violence between individuals who are unrelated, and who may or may not know each another, generally taking place outside the home (WHO, 2002). In some communities, due to high violence, low neighborhood cohesion, and elevated social disorganization, gang activity maybe prevalent (Brody et al., 2001; Rankin & Quane, 2002). Unfortunately, acceptance into a group such as a gang depends on adhering to group norms that may sanction not only delinquent behaviors but also health risk behaviors (Molidor, 1996). Therefore, it is not unexpected that youth held in detention represent a highly vulnerable population for engaging in illicit drug use and risky sexual behaviors (Crosby, DiClemente, Staples-Horne, 2003; Teplin et al., 2002).

Detained youth report higher rates of marijuana, amphetamines, and crack use compared to other adolescent groups (Eng and Butler, 1997). In addition, recent evidence from the Centers for Disease Control and Prevention indicated that rates of gonorrhea among adolescent detainees were almost 7 and 10 times greater among these males and females, respectively, when compared to adolescents from the general population (Center for Disease Control and Prevention, 2002). Furthermore, surveillance surveys have consistently shown that detained adolescents have high rates of chlamydia and gonorrhea infection than other adolescent groups (Canterbury et al., 1995; Pack, DiClemente, Hook & Oh, 2000). Unfortunately, the pathways to such behaviors are not sufficiently understood.

A growing body of literature documents significant associations between community violence exposure and drug and sexual risk behaviors. Collectively, these studies document that youth who are exposed to high rates of violence within their communities are at greater risk for engaging in behaviors such as illicit drug use and unsafe sexual behaviors (Berenson, Wiemann, & McCombs, 2001; Stiffman, Dore, Cunningham, & Earls, 1995; Voisin, 2003, 2005). However, the pathways which mediate such relationships are not adequately understood or empirically researched. Given that gang involvement is also associated with drug use and unsafe sex (Harper & Robinson, 1999; Huff 1998; Thornberry & Burch, 1997; Wingood et al., 2002; Voisin et al., 2004), it may provide a mechanism through which community violence exposure is associated with such youth risk outcomes. Such a mechanism may be widely speculated in clinical and community settings. However, few studies have empirically examined this assumption. Furthermore, no studies to date, have examined the interrelationships among these variables with detained youth. This is unfortunate give the disproportionate vulnerabilities of these youth with regard to these factors. A final contribution of this study is that we offer a theoretical explanation of how peer influences may mediate the relationship between community violence exposure and drug and sexual risks.

The conceptual framework for this study is informed by social control theory (SCT) (Elliott et al., 1985; 1989; Hirschi, 1969). According to one application of this theory, high rates of community violence result in social disorganization, which weakens the ability of prosocial institutions (e.g., schools, families) to control and monitor their youth (Farrington et al., 1990). Consequently, failed schools and local social institutions offer adolescents scant hope for the future, and adolescents feel uncommitted to conventional societal norms (Petraitis, Flay & Miller, 1995). Weak attachments to prosocial institutions, which generally oppose drug and sexual risk behaviors, may heighten the risk of adolescents becoming more attached to deviant peers who in turn may promote risky norms resulting in increased drug use and increased sexual risk behaviors (Petraitis, Flay & Miller, 1995).

Specifically, this study sought to test three hypotheses: 1) increased exposure to community violence will be associated with heightened drug and sexual risk behaviors; 2) higher levels of community violence exposure will be associated with gang involvement and perceived negative peer norms; and 3) gang involvement and perceived negative peer norms will mediate the relationship between community violence exposure and drug and sexual risk behaviors.

Prior research has documented that a number of contextual factors are associated with drug and sexual risk behaviors among adolescents. Gender, ethnicity, and socioeconomic status have all been shown to be significantly correlated to both exposures to community violence (Boney-McCoy & Finkelhor, 1995; Fitzpartick & Boldizar, 1993) and drug and sexual risk behaviors (Canterbury et al., 1995; Rickman et al., 1994). Furthermore, previous studies have hypothesized that contextual effects such as neighborhood disadvantage on delinquent behaviors have been explained partially by parenting behaviors (Brody et al., 2001; Conger, Ge, Elder, Lorenz, & Simmons, 1994; McLoyd, 1990). Along these lines, other studies have shown associations between adverse sexual and drug risk behaviors and low familial social support (Resnick et al., 1997; Voisin, 2002), low parental monitoring (Sheidow, Gorman-Smith, Tolan & Henry, 2001) and family violence exposure (Voisin, 2005). More recently, research has documented an inverse relationship between religiosity and delinquency among youth (Regnerus, 2003). In testing all hypotheses we controlled for the potential effects of these covariates.

METHODS

Subject Population

Between October 2001 and July 2003, youth were recruited from 8 regional youth detention centers (RYDCs) across the state of Georgia. At that time in Georgia, the male to female ratio of detained youth was approximately 3 to 1 with many RYDCs serving a maximum of 30 youth. The 8 RYDCs selected to participate in this study had a higher maximum capacity than 30 and also maintained a < 3:1 ratio of males to females; thus, there were more detained females from which to sample. This sampling strategy was chosen to obtain a balanced number of females in the overall sample. Given that we were not permitted to utilize a sampling frame and that the State Department of Juvenile Justice’s Internal Review Board (IRB) required us to obtain parental consent, this study employed a convenience sample.

Participants were detained for a variety of reasons such as property and personal crimes, curfew and truancy violations, vandalism, and allegations of sexual abuse. Some participants were serving a short-term sentence (i.e., < 60 days) of incarceration. We obtained permission to collect data from participants on the terms that we would not collect data on individual offense status.

The study achieved an 82% participation rate. Recruitment efforts occurred weekly at each of the eight facilities. Research assistants recruited only newly-admitted adolescents. Adolescents were eligible if they were 14 to 18 years old, expressed willingness to participate, and had a parent or legal guardian provide informed consent. By agreement with both the Georgia Department of Juvenile Justice’s Internal Review Board (IRB) and with Emory University’s IRB, research assistants were permitted to obtain parental/legal guardian consent verbally by phone. Both IRBs approved all study procedures.

Data Collection

All self-reported measures were assessed using audio-computer assisted self-interviewing (A-CASI), based on findings suggesting decreased reporting bias compared to paper and pencil approaches (Michaud, Narring, & Ferron, 1999). By providing a voice track that delivered each question to adolescents through headphones, A-CASI technology may have reduced problems that would have otherwise been posed by illiteracy. The technology also creates a user-friendly interview method that automatically handles skip patterns in the questionnaire and provides adolescents with an interactive experience (possibly increasing task attentiveness and completion). Adolescents’ responses to the computer-delivered questions were automatically encrypted to ensure confidentiality which may have reduced response bias.

Variables of Interest

Age

Age was assessed by one item: “How old are you?” (range 14 to 18 years).

Ethnicity

Ethnicity was categorized as 0=White, 1= African American (Black), 3=Hispanic, and 4=all other ethnic groups. To ensure sufficient cell sizes for the path analyses reported below, this variable was collapsed into a dichotomous exogenous variable in those analyses (White versus Non-white).

Gender

Gender was assessed by one item: “What is your gender?” (1= male and 2=female).

Socioeconomic Status

Eligibility for free school lunch, a proxy for socioeconomic status was measured by one item: “Do you or would you qualify for a free lunch at school?” (No-Yes).

Religion

Religiosity was measured by a 7-item scale (Regnerus, 2003). Seven behaviors were presented to adolescents, and they were asked to indicate how often in a typical month they engaged in the following: attend church or religious services, pray or meditate, talk to others about religious or spiritual concerns, talk with a religious leader such as minister or priest, watch religious TV programs, read your bible or other religious material, and say prayers or grace before meals in your home. The response format ranged from 1 (never) to 4 (very often). Cronbach’s alpha = .84.

Parental Monitoring

General level of monitoring by parental figure(s) was assessed by a 6-item scale that assessed adolescent perception of parental knowledge of their whereabouts as well as disclosure of their activities (DiClemente, et al., 2001) (e.g., “When you are away from home and not at school or work, does this person know where you are?”). Participants indicated their level of agreement or disagreement using a 5-point Likert-type response scale with each item ranging from 1 (never) to 5 (always). Cronbach’s alpha = .88.

Familial Social Support

General familial social support was measured by a 4-item family subscale from the Multidimensional Scale of Perceived Social Support (Zimet, Dahlem, Zimet, & Farley, 1988) (e.g. “My family really tries to help me”). Adolescents indicated their level of agreement or disagreement using a 5-point Likert-type response scale with each item ranging from 1 (strongly agree) to 5 (strongly disagree). Cronbach’s alpha = .91.

Family Violence

Lifetime exposure to family violence was assessed by 8 items (e.g., how many times have you seen or heard your mom’s partner hurt her physically? By hurt I mean slapping, punching, choking or kicking her? Adolescents indicated their level of agreement or disagreement using a 5-point Likert-type response scale with each item ranging from 1 (never) to 5 (always). Cronbach’s alpha = .90.

Community Violence

Exposure to community violence in the past 12 months, prior to being detained was assessed by 6 items from the community violence subscale from the stressful life events index (Attar, Guerra, & Tolan, 1994). 1) Has a family member been robbed or attacked? 2) Has someone else you know other than a family member gotten beaten up, attacked, or really hurt by others? 3) Have you seen anyone beat, shot, or really hurt by someone? 4) Have you been around people shooting guns? 5) Have you been afraid to go outside to play, or have your parents made you stay inside because of guns or drugs in your neighborhood? 6) Have you had to hide someplace because of shootings in your neighborhood? The response format was (0) No or (1) Yes. A violence exposure score was computed for each participant by summing responses which represented a range of exposure to various types of community violence. Cronbach’s alpha = .73.

Gang membership

Lifetime gang membership was assessed by one item, “Have you ever been a member of a gang?” Current gang membership was assessed by a single item, “Are you currently a member of a gang?” (No-Yes).

Perceived Risky Peer Norms

Perceived risky peer norms were measured by a 5-item scale (DiClemente et al., 2001) (e.g., “Most of the girls/guys I know never use condoms when they have sex”). Participants responded on a 4-point Likert type scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Cronbach’s alpha = .86)

Substance Use

Five separate items assessed incidence of drug-use behaviors in the 2 months prior to being detained. Participants were asked to answer 0 (No) or 1 (Yes) to whether they had used marijuana, amphetamines, crack, ecstasy, or alcohol. A Drug-use index was calculated from the total scores for each participant.

HIV Risk Behaviors

Six separate items assessed STD risk behaviors over the last 2 months prior to being detained. Sex was defined as having either oral or vaginal sex with the opposite sex. Participants were asked whether they had sex without a condom during the time of last sex, and in the prior 2 months, had sex while high on alcohol or drugs, had sex with a partner who was high on alcohol or drugs, had sex with 2 or more people at the same time, traded sex for drugs, and traded sex for money. The response format was (0) No or (1) Yes. An HIV risk index was calculated from the overall score.

Statistical Analyses

Initial analyses profiled the characteristics of the sample via one-way frequency tables and measures of central tendency. Following initial univariate analyses, based on the hypotheses articulated above we fit a path model to the data in which we tested direct paths from community violence exposure to lifetime gang membership, perceived risk norms, sexual risk behaviors and substance use after controlling for myriad contextual influences (i.e., family social support, family violence, receipt of welfare, parental monitoring, perceived influence of religion in one’s life) and demographic variables (age, gender, and non-white ethnicity). The model also tested direct paths from lifetime gang membership to sexual risk behaviors and substance use; direct paths from endorsement of perceived risk norms to sexual risk behaviors and substance use; and the indirect paths from community violence exposure to sexual risk behaviors and substance use through gang involvement and through perceived risk norms. A conceptual diagram of the path analysis model is shown in Figure 1.

Figure 1.

Figure 1

Open in a new tab

Conceptual Path Diagram of Relationships between Exogenous Explanatory Variables, Gang Membership and Risk Norm Mediators, and Sexual Risk and Substance Use Outcome Variables.

Because our constructs were measured by single indicators (e.g., gender, age, lifetime gang membership) or previously validated scales (e.g., parental monitoring, community violence), we treated these variables as observed rather than latent. Given the lack of research to date investigating the impact of community violence on HIV risk behaviors, we fit a fully-specified (i.e., saturated) path analysis model to illuminate all possible direct and indirect relationships among exogenous, intermediary, and outcome variables.

We conducted the path analysis using Mplus 3.13 for Windows (Muthén & Muthén, 2004). Due to the dichotomous nature of lifetime gang membership, we used weighted least-squares estimation with a mean and variance adjustment (Mplus estimator WLSMV). We computed unstandardized regression coefficients and 95% confidence intervals for each direct and indirect effect in the analysis, as well as the corresponding standardized regression coefficients. Due to the presence of indirect effects in the analysis, we computed confidence intervals via the bias-corrected (BC) bootstrap for non-symmetric indirect effect distributions, as recommended by Shrout & Bolger (2002) and MacKinnon et al. (2004). The number of bootstrap samples was set above 5,000 to ensure sufficient precision of the confidence intervals (Hox, 2002). Confidence intervals that do not include zero imply that the parameter estimate around which the confidence interval is constructed is statistically significant at p < .05.

RESULTS

Sample Characteristics

The overall study sample comprised 559 adolescents. Of the sample, half were female (50.1%), 46.7% identified as White, 46.0% as Black, and 7.3% as Other. The average age of these adolescents was 15.4 years (SD = .96), and the median level of education was completion of the 9th grade. Most respondents were between 14-15 years old (32.7%) or 16-17 years old (45.4%). Approximately one-quarter (25.6%) of adolescents lived with both parents, while almost half (45.1%) lived with their mothers only. A substantial minority of respondents lived with a relative other than their mothers or fathers (19.6%) while 7.3% lived with their fathers only. The majority of adolescents were eligible for free school lunches (57.5%). Adolescents had been serving their current detention sentence for a median of 14 days, and slightly more than two-thirds (68%) reported serving at least one detention sentence previously.

Community Violence Exposure

A large majority (78.9%) of these youth reported exposure to one or more incidents of community violence within the past 12 months, prior to being detained. More than two-thirds, (76.3%) had witnessed one or more types of violence exposure in the last 12 months. Of those witnessing violence, 19.7% had a family member robbed or attacked; 37.6% had seen someone other than a family member beaten or attacked by others; 56.1% had seen someone beaten, shot, or really hurt by someone; 52.5% had been around people shooting guns in their neighborhood; 11.5% had been afraid to go outside and play, or their parents made them stay inside because of gangs or drugs in their neighborhood; and 11.3% had to hide somewhere because of shootings in their neighborhood.

Risk Behaviors

Of the total sample, 27.3% ever belonged to a gang. Of those who had ever belonged to a gang, 52.9% currently belonged to a gang. In terms of drug use, 50.4% smoked marijuana, 56.5% used amphetamines, 30.6% crack, 57.1% ecstasy, and 73% alcohol in the 2 months before coming to the center. Among sexually-active adolescents (n=468), 58.3% did not use male condoms during last sex, 4.3% traded sex for drugs, 56.0% had sex while high on alcohol or drugs, 52.9% had sex with a partner who was high on alcohol or other drugs, and 23.1% had sex concurrently with two or more persons, within the 2 months before coming to the center.

Relationships among Variables

Correlations among the measured variables showed that experiences of community violence were positively associated with sexual risk behavior (r = .17, p < .0001) and substance use (r = .12, p < .05). Lifetime history of gang membership was positively correlated with both sexual risk (r = .14, p < .05) and substance use (r = .11, p < .05). Endorsement of perceived risk norms was also positively related to sexual risk (r = .17, p < .001) and substance use (r = .09, p < .05). Increased levels of parental monitoring were associated in a lower likelihood of sexual risk (r = −.10, p < .05) and substance use (r = −.20, p < .0001).

Path Analysis Results

Unstandardized regression coefficients and 95% confidence intervals for direct effects appear in Table 2.

Table 2.

Unstandardized Regression Coefficients (95% Confidence Intervals) for Direct Effects

Outcome Variable Gang Membership Risk Norms Sexual Risk Substance Use
Explanatory Variable
  Community Violence 0.17 (0.06, 0.28) 0.20 (−0.31, 0.69) 0.09 (0.01, 0.19) 0.11 (−0.002, 0.22)
  Family Social Support 0.01 (−0.03, 0.05) 0.06 (−0.10, 0.22) −0.01 (−0.04, 0.03) 0.02 (−0.02, 0.06)
  Family Violence 0.19 (−0.11, 0.44) 0.90 (−0.30, 2.14) −0.04 (−0.26, 0.16) −0.21 (−0.48, 0.05)
  Welfare −0.01 (−0.51, 0.40) 0.79 (−0.85, 2.21) −0.09 (−0.43, 0.25) 0.004 (−0.43, 0.38)
  Gender −0.08 (−0.37, 0.22) 1.68 (0.52, 2.78) −0.04 (−0.27, 0.22) 0.25 (−0.04, 0.55)
  Age −0.10 (−0.25, 0.05) 0.04 (−0.58, 0.61) −0.04 (−0.16, 0.07) 0.19 (0.05, 0.33)
  Non-White Ethnicity 0.22 (0.01, 0.44) 0.17 (−0.65, 0.98) −0.13 (−0.28, 0.04) −0.66 (−0.86, −0.43)
  Parental Monitoring −0.04 (−0.07,−0.01) −0.02 (−0.16, 0.10) −0.01 (−0.04, 0.02) −0.04 (−0.07, −0.01)
  Religious Influence 0.01 (−0.03, 0.05) −0.001 (−0.15, 0.15) 0.01 (−0.02, 0.04) −0.03 (−0.06, 0.01)
  Gang Membership — 1.49 (0.74, 2.20) 0.11 (−0.06, 0.26) 0.23 (0.05, 0.43)
  Risk Norms — — 0.03 (0.01, 0.06) 0.01 (−0.02, 0.04)
Open in a new tab

N = 559.

Confidence intervals are computed via the bias-corrected bootstrap based on 5,001 bootstrap samples. Boldface type indicates confidence intervals that do not include zero.

The corresponding standardized direct effects appear in Table 3.

Table 3.

Standardized Regression Coefficients for Direct Effects

Outcome Variable Gang Membership Risk Norms Sexual Risk Substance Use
Explanatory Variable
  Community Violence .23 .05 .12 .14
  Family Social Support .04 .04 −.03 .07
  Family Violence .13 .12 −.03 −.13
  Welfare −.01 .14 −.08 .003
  Gender −.07 .29 −.04 .20
  Age −.08 .01 −.03 .14
  Non-White Ethnicity .20 .03 −.11 −.53
  Parental Monitoring −.23 −.02 −.06 −.18
  Religious Influence .05 −.001 −.03 −.12
  Gang Membership — .29 .10 .21
  Risk Norms — — .16 .06
Open in a new tab

N = 559.

Boldface type indicates 95% confidence intervals for corresponding unstandardized coefficients that do not include zero, as shown in table 2.

As shown in Table 3, the following direct effects were found to be significant: Community violence was positively associated with lifetime gang membership (β = .23) and risky sexual behavior (β =.12). Non-white ethnicity was also positively associated with lifetime gang membership (β = .20). Higher levels of parental monitoring resulted in a lower likelihood of lifetime gang membership (β = −.23). Being female was positively associated with endorsement of higher perceived risk norms (β = .29) as was a prior history of gang membership (β = .29). Perceived risky peer norms was in turn associated within increased sexual risk behavior (β = .16). Being older (β = .14), and a history of gang membership (β = .21) were positive correlates of substance use. However, higher levels of parental monitoring exerted a protective effect by lowering the likelihood of substance use (β = −.18). Finally, non-white race was negatively associated with substance use (β = −.53).

Direct effects examine the most immediate impact of a hypothesized explanatory variable such as community violence on distal outcomes (e.g., sexual risk behavior; substance use). This path analysis model has two intermediary variables, gang membership and perceived risky peer norms that we hypothesized to mediate the relationship between community violence exposure and drug and sexual risk behaviors. Based on our model, community violence did have an indirect relationship to drug and sexual risk behaviors through three possible pathways as shown in Figure 1: (a) gang membership alone; (b) risk norms alone, and (c) gang membership followed by risk norms.

Mediating role of peer influences

We found a significant indirect relationship between community violence exposure and substance use through gang membership alone (B = 0.040; 95% CI = 0.008, 0.096; β = .355). We also found that similar indirect relationship for non-white ethnicity (B = 0.052; 95% CI = 0.002, 0.163; β = .166), and parental monitoring (B = −0.010; 95% CI = −0.025, −0.001; β = −.288). These findings suggest that community violence exposure and non-white ethnicity is associated with increased risk for substance use in part through gang membership. Whereas, parental monitoring is associated with lower levels of substance use through reduced gang membership.

For sexual risk behavior, we found significant indirect relationships for community violence exposure (B = 0.008; 95% CI = 0.002, 0.023; β = 1.085) and non-white ethnicity (B = 0.011; 95% CI = 0.001, 0.034; β = 1.114) through the path of gang membership to risk norms. This suggests that community violence exposure and non-white ethnicity are both significant correlates to gang membership, which in turn is associated with endorsement of more risky norms. Consequently, the likelihood of adolescents reporting risky sexual behavior was increased. No significant indirect effects were found when risk norms were considered as a single mediating variable.

DISCUSSION AND IMPLICATIONS

Using a heterogeneous sample of detained youth, this study examined the relationship between community violence exposure and drug and sexual risk behaviors, and explored for potential mediating mechanisms that could explain the relationships. Levels of community violence in this population were high, yet consistent with levels of exposure reported among other study populations such as urban youth (Bell & Jenkins 1993; Sheidow, Gorman-Smith, Tolan, & Henry, 2001; Voisin, 2002). Findings indicated that community violence exposure was significantly associated with drug and sexual risk behaviors among detained youth. These findings corroborate those of other studies utilizing community-based samples of youth (Berenson, Wiemann, & McCombs, 2001; Stiffman, Dore, Cunningham, & Earls, 1995; Voisin, 2003, 2005). Taken together, these results suggest the need to develop better assessment instruments to identify youth who are exposed to both community violence and engage in drug and sexual risk behaviors. Identifying drug and sexual risks among detained youth during assessments should prompt clinicians and other service providers to screen for the presence of community violence, and vice-versa.

Unlike prior research in this area, this study examined not only the direct relationship between community violence and health-risk behaviors, but also the indirect effects through two theoretically-derived factors—gang membership and negative peer influences. As we speculated, and as the results indicate, exposure to community violence had significant direct relationship to both lifetime gang membership and sexual risk behavior. Exploring the pathways from community violence exposure to these risk behaviors, we found several notable indirect pathways: one was the indirect pathway to sexual risk behavior through gang membership and perceived risky peer norms; and a second one was on substance use through gang membership, and perceived risky norms. Findings suggest that a heightened level of exposure to community violence even after controlling for other myriad environmental influences is associated with an increased gang membership which is correlated with perceived risky peer norms and sexual risk behaviors and substance use. These findings suggest that multiple intervention points may be possible for disrupting the potential pathways to drug and sexual risk among detained youth. Interventions designed to alter negative peer affiliations and promote membership in prosocial peer groups might be a useful strategy. Another approach is to use former gang members in group interventions with current gang members to alter the perception of risky peer norms.

Clearly, youth who belong to gangs endorse social norms that are more supportive of engaging in risky sexual behaviors such as not using condoms, having multiple sex partners, and using substances while having sex. For example, research with gang members has documented gang initiation rites that involve extreme sexual practices such as “pulling a train” (i.e., having sex with multiple male gang members) (Molidor, 1996). What is not clear is what factors related to community violence exposure motivates these youth to join a gang. In particular, we found that females and Blacks were more likely to be gang members than males and whites, and females were more likely to perceive their peers endorsing risky norms than males. Research with female adolescents suggests that peer influences, dysfunctional families and living in neighborhoods characterized by low educational opportunities, violence, and poverty play a role in female gang involvement (Molidor, 1996; Walker & Mason, 2001). Other research has shown that among African American populations, environmental factors such as the presence of drug dealers in the neighborhood, having peers who use drugs, and having a family member who belongs to a gang were predictors of gang membership (Curry & Spergel, 1992). It may be that among our sample of detained youth, especially the females and the Black adolescents, they are trying to feel a sense of belonging and power by joining a gang because they are afraid, or they may be seeking protection from the violence they witness in their community. Some research also suggests that belonging to a gang provokes respect and status among marginalized groups (Walker-Barnes & Mason, 2001). Thus, in communities plagued by high levels of violence, respect may equate with safety as they are proffered some level of protection through their gang membership. There has been some research highlighting the interrelationship of gang involvement, delinquency and social contexts (e.g., Curry & Spergel, 1992); however, future research should investigate these specific underlying motivating factors among this population to understand their reasons for gang involvement so that intervention efforts can be developed. Moreover, these studies should examine the differential gender and racial/ethnic factors involved so that efforts can be tailored. Community violence appears to be linked to sexual risk behavior via gang membership and endorsement of risk norms. This study found that detained females in this sample were more likely than males to perceive more risky peer norms. These findings might suggest that girls who are detained might be more disturbed than their male counterparts.

Several protective factors were significant at various risk junctures. Increased perceived parental monitoring was negatively associated with both gang involvement and substance use among youth in this sample. Training parents to adequately monitor their youth, or assisting them with networking with other parents or adults who can provide monitoring functions for their youth may be one effective strategy for disrupting the likelihood of joining gangs or abusing substances.

This is the first study to examine pathways between community violence exposure and HIV-related risk behaviors and drug and substance use among detained youth. Although the findings are informative they must be interpreted within the methodological limitations of the study. Foremost the cross sectional nature of the study precludes assumptions of causality and we can only evaluate associations. However, it should be noted that in this context true experimental designs cannot be executed to determine causality (i.e., persons cannot be randomly assigned to violent versus non violent conditions). Directionality between some variables is tentative. For instance, we assessed community violence exposure and drug and sexual risk behaviors at 12-month and 2-month time frames respectively. Such temporal ordering may imply directionality. However, with regards to community violence exposure and gang involvement the temporal ordering of the relationship is unclear. For instance, gang involvement may lead to increased community violence exposure, or adolescents may join gangs as a consequence of exposure to such violence. Although some researchers have found that community violence exposure is an antecedent to gang involvement (Tolan, Gorman-Smith, & Henry, 2003), future studies using longitudinal designs will need to further disentangle such associations. Findings may also be limited by the validity of the self-reported measures, by the use of a convenience sample of detained adolescents in part necessitated by the need to obtain parental consent and by the possibility of response bias, although sophisticated computer technology was employed in order to reduce such occurrences. Path analysis was very general, which was appropriate at this early stage of an exploratory study. In spite of the above limitations, common in all cross-sectional research, we believe that the significant predictors we found in this study will aid researchers in designing longitudinal observational studies to advance research in this important area.

Table 1. Mechanisms linking violence exposures and school engagement among African American adolescents: Examining the roles of psychological problem behaviors and gender.

Correlations among observed variables

1 2 3 4 5 6 7 8 9 10 11 12
1. Community Violence 1.00
2. Family Social Support −.09* 1.00
3. Family Violence .24**** −.18*** 1.00
4. Welfare .05 −.12** −.03 1.00
5. Gendera −.10* −.17*** .20**** .14** 1.00
6. Age .05 −.06 .01 −.01 −.06 1.00
7. Non-White Ethnicityb .14 −.02 −.02 .08 −.01 −.03 1.00
8. Parental Monitoring −.25**** .38**** −.06 −.03 .20**** −.18*** −.11* 1.00
9. Religious Influence −.06 .20**** .004 .11* .04 −.05 .03 .23**** 1.00
10. Gang Membershipc .27**** −.05 .13** .02 −.08 −.02 . 15*** −.06 .001 1.00
11. Risk Norms .17*** −.11* .24**** .13** .25**** −.02 .10* −.23**** .04 .24**** 1.00
12. Sexual Risk .17*** −.04 .08 −.02 −.04 −.02 −.04 −.10* −.01 .14* .17*** 1.00
13. Substance use .12* −.07 .01 −.01 .05 .16*** −.33**** −.20**** −.16*** .11* .09* .09*
Open in a new tab

N = 559.

a

Gender was coded as 1 = male; 2 = female.

b

Ethnicity was coded as 0 = white; 1 = non-white.

c

Gang membership was coded as 0 = no lifetime history of gang membership; 1 = lifetime history of gang membership. Correlations were estimated using full information maximum likelihood in Mplus 3.13.

Acknowledgements

This research was supported, in part, by the Emory Center for AIDS Research (NIH/NIAID 2 P30 AI50409-04A1), the Rural Center for AIDS/STD Prevention at Indiana University, a grant from the University Research Council at Emory University, and by a grant award to the Center for AIDS Prevention Studies R25 HD045810-02.

References

  1. Attar BK, Guerra NG, Tolan PH. Neighborhood disadvantage, stressful events, and adjustment in urban elementary school children. Journal of Clinical Child Psychology. 1994;23:391–400. [Google Scholar]
  2. Bell C, Jenkins E. Community violence and children on Chicago’s Southside. Psychiatry. 1993;56:46–54. doi: 10.1080/00332747.1993.11024620. [DOI] [PubMed] [Google Scholar]
  3. Berenson A, Wiemann C, McCombs S. Exposure to violence and associated health-risk behaviors among adolescent girls. Achieves of Pediatrics and Adolescent Medicine. 2001;15:1238–1242. doi: 10.1001/archpedi.155.11.1238. [DOI] [PubMed] [Google Scholar]
  4. Boney-McCoy S, Finkelhor D. Psychosocial sequelae of violent victimization in a national youth sample. Journal of Consulting and Clinical Psychology. 1995;63:726–736. doi: 10.1037//0022-006x.63.5.726. [DOI] [PubMed] [Google Scholar]
  5. Brody GH, Ge X, Conger R, Gibbons FX, Murry VM, Gerrard M, Simons RL. The influence of neighborhood disadvantage, collective socialization, and parenting on African American children’s affiliation with deviant peers. Child Development. 2001;72:1231–1246. doi: 10.1111/1467-8624.00344. [DOI] [PubMed] [Google Scholar]
  6. Canterbury R, McGarvey E, Sheldon-Keller A, Waite D, Reams P, Koopman C. Prevalence of HIV-related risk behaviors and STDs among incarcerated adolescents. Journal of Adolescent Health. 1995;17:173–177. doi: 10.1016/1054-139X(95)00043-R. [DOI] [PubMed] [Google Scholar]
  7. Centers for Disease Control . HIV/AIDS Surveillance-June 2002. Dept. of Health and Human Services, Public Health Service, Centers for Disease Control; Atlanta, GA: 2002. [Google Scholar]
  8. Chung HL, Steinberg L. Relations between neighborhood factors, parenting behaviors, peer deviance, and delinquency among serious juvenile offenders. Developmental Psychology. 2006;42(2):319–331. doi: 10.1037/0012-1649.42.2.319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Conger R, Ge X, Elder GH, Lorenz FO, Simons RL. Economic stress, coercive family process and development problems of adolescents. Child Development. 1994;65:541–561. [PubMed] [Google Scholar]
  10. Crosby R, DiClemente R, Staples-Horne M. Health issues of juvenile offenders. In: Moore J, editor. Management and administration of correctional healthcare. Civic Research Institute; Kingston NJ: 2003. pp. 11–12. [Google Scholar]
  11. Curry G, Spergel I. Gang involvement among Hispanic and African American adolescents. Crime and Delinquency. 1992;29(3):273–291. [Google Scholar]
  12. DiClemente RJ, Wingood GM, Crosby RA, Cobb BK, Harrington K, Davies S, Hook EW, Oh MK. Parental monitoring and its association with a spectrum of adolescent health risk behaviors. Pediatrics. 2001;107:1363–1368. doi: 10.1542/peds.107.6.1363. [DOI] [PubMed] [Google Scholar]
  13. Elliott D, Huizinga D, Ageton S. Explaining delinquency and drug use. Sage; Beverly Hills, CA: 1985. [Google Scholar]
  14. Elliott D, Huizinga D, Manard S. Multiple problem youth: delinquency substance use and mental health problems. Springer-Verlag; New York: 1989. [Google Scholar]
  15. Eng T, Butler W. The hidden epidemic: Confronting sexually transmitted diseases. National Academy Press; Washington, DC: 1997. [PubMed] [Google Scholar]
  16. Farrington D, Loeber R, Elliott D, Hawkins d., Kandel D, Klein M, McCord J, Roew D, Tremblay R. Advancing knowledge about the onset of delinquency and crime. Advances in Clinical Child Psychology. 1990;13:283–342. [Google Scholar]
  17. Fitzpatrick K, Boldizar J. The prevalence and consequence of exposure to violence among African-American youth. Journal of the American Academy of Child and Adolescent Psychiatry. 1993;32:424–430. doi: 10.1097/00004583-199303000-00026. [DOI] [PubMed] [Google Scholar]
  18. Harper GW, Robinson WL. Pathways to risk among inner-city African-American adolescent females: The influence of gang membership. American Journal of community Psychology. 1999;27(3):383–404. doi: 10.1023/A:1022234027028. [DOI] [PubMed] [Google Scholar]
  19. Hill KG, Howell JC, Hawkins JD, Battin-Pearson SR. Childhood risk factors for adolescent gang membership: Results from the Seattle Social Development Project. Journal of Research in Crime and Delinquency. 1999;36(3):300–322. [Google Scholar]
  20. Hirschi T. Causes of delinquency. University of California Press; Berkeley, CA: 1969. [Google Scholar]
  21. Hox JJ. Multilevel Analysis: Techniques and Applications. Lawrence Erlbaum and Associates; Mahwah, New Jersey: 2002. [Google Scholar]
  22. Huff CR. Criminal behavior of gang members and at-risk youths. U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention; Washington, DC: 1998. NCJ Publication No. 164725. [Google Scholar]
  23. Jessor R, Donovan JE, Windmer K. Psychosocial factors in adolescent alcohol and drug use: The 1980 national sample study and the 1974-78 panel study. University of Colorado, Institute of Behavioral Science; Boulder: 1980. Unpublished final report. [Google Scholar]
  24. MacKinnon DP, Lockwood C. Confidence Limits for the Indirect Effect: Distribution of the Product and Resampling Methods. Multivariate Behavioral Research. 2004;39(1):99–128. doi: 10.1207/s15327906mbr3901_4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. McLoyd VC. The impact of economic hardship on Black families and children: Psychological distress, parenting, and socioemotional development. Child Development. 1990;61:311–346. doi: 10.1111/j.1467-8624.1990.tb02781.x. [DOI] [PubMed] [Google Scholar]
  26. Margolin G, Gordis E. The effects of family and community violence on children. Annual Review of Psychology. 2000;51:445–479. doi: 10.1146/annurev.psych.51.1.445. [DOI] [PubMed] [Google Scholar]
  27. Margolin G. Effects of domestic violence on children. In: Trickett PK, Schellenback CJ, editors. Violence against children in the family and the community. American Psychiatric Association; Washington, DC: 1998. pp. 57–102. [Google Scholar]
  28. Michaud P, Narring F, Ferron C. Alternative methods in the investigation of adolescents’ sexual life. Journal of Adolescent Health. 1999;25:84–90. doi: 10.1016/s1054-139x(98)00158-x. [DOI] [PubMed] [Google Scholar]
  29. Molidor C. Female gang members: a profile of aggression and victimization. Social Work. 1996;41(3):251–257. [PubMed] [Google Scholar]
  30. Muthén LK, Muthén B. Mplus User’s Guide. Muthen and Muthen, Inc.; Los Angeles, CA: 2004. [Google Scholar]
  31. Ng-Mak DS, Salzinger S, Feldman R, Stueve A. Normalization of violence among inner-city youth: A formulation for research. American Journal of Orthopsychiatry. 2002;72:92–101. [PubMed] [Google Scholar]
  32. Pack P, DiClemente R, Hook E, Oh M. High prevalence of asymptomatic STDs in incarcerated minority male youth: A case for screening. Sexually Transmitted Diseases. 2000;27:175–177. doi: 10.1097/00007435-200003000-00011. [DOI] [PubMed] [Google Scholar]
  33. Petraitis J, Flay B, Miller T. Reviewing theories of adolescent substance abuse: Organizing pieces of the puzzle. Psychological Bulletin. 1995;117(1):67–86. doi: 10.1037/0033-2909.117.1.67. [DOI] [PubMed] [Google Scholar]
  34. Puzzanchera C, Stahl A, Finnegan T, Tierney N, Snyder H. Juvenile Court Statistics. U.S. Department of Justice. Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention; Washington, DC: 2003. [Google Scholar]
  35. Rankin BH, Quane JM. Social contexts and urban adolescent outcomes: The interrelated effects of neighborhoods, families, and peers on African-American youth. Social Problems. 2002;49:79–100. [Google Scholar]
  36. Regnerus M. Moral communities and adolescent delinquency: religious contexts and community social control. Sociological Quarterly. 2003;44:523–554. [Google Scholar]
  37. Resnick M, Bearman PS, Blum RW, Bauman KE, Harris KM, Jones J, Tabor J, Beuhring R, Sieving RE, Shew M, Ireland M, Bearinger LH, Udry JR. Protecting adolescents from harm: Findings from the National Longitudinal Study on Adolescent Health. Journal of American Medical Association. 1997;278(10):823–832. doi: 10.1001/jama.278.10.823. [DOI] [PubMed] [Google Scholar]
  38. Rickman R, Lodico M, DiClemente R, Morris R, Baker C, Huscroft S. Sexual communication is associated with condom use by sexually active incarcerated youth. Journal of Adolescent Health. 1994;15(5):383–388. doi: 10.1016/1054-139x(94)90261-5. l. [DOI] [PubMed] [Google Scholar]
  39. Sheidow J, Gorman-Smith D, Tolan P, Henry D. Family and community characteristics: risk factors for violence exposure in inner-city youth. Journal of Community Psychology. 2001;29:345–360. [Google Scholar]
  40. Shrout PE, Bolger N. Mediation in Experimental and Nonexperimental Studies: New Procedures and Recommendations. Psychological Methods. 2002;7(4):422–445. [PubMed] [Google Scholar]
  41. Sickmaund M, Sladky T, Kang W. [Accessed June 6, 2006];Census of Juveniles in Residential Placement Data book. 2003 Available at: http://ojjdp.ncjrs.org/ojstabb/cjrb.
  42. Snyder H. Juvenile Arrests 2003. Office of Juvenile and Delinquency Protection. Office of Juvenile Justice and Delinquency Protection. OJJDP Statistical Briefing Book; Washington, DC: 2003. Available at: http://ojjdp.ncjrs.org/ojstabb/crime/qa05101. [Google Scholar]
  43. Stiffman A, Dore P, Cunningham R, Earls F. Person and environment in HIV risk behavior change between adolescence and young adulthood. Health Education Quarterly. 1995;22(2):211–226. doi: 10.1177/109019819502200209. [DOI] [PubMed] [Google Scholar]
  44. Teplin L, Mericle A, McClelland G, Abram K. HIV and AIDS Risk Behaviors in Juvenile Detainees: Implications for Public Health Policy. American Journal of Public Health. 2002;93(6):906–912. doi: 10.2105/ajph.93.6.906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Thornberry TP, Burch JH. Gang members and delinquent behavior. U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention; Washington, DC: 1997. NCJ Publication No. 165154. [Google Scholar]
  46. Tolan P, Gorman-Smith S, Henry D. The developmental ecology of urban males’ youth violence. Developmental psychology. 2003;39(2):274–291. doi: 10.1037//0012-1649.39.2.274. [DOI] [PubMed] [Google Scholar]
  47. U.S. Bureau of the Census . March current population survey. U.S. Government Printing Office; Washington, DC: 2000. [Google Scholar]
  48. Voisin D. Family ecology and HIV sexual risk behaviors among African American and Puerto Rican adolescent males. American Journal of Orthopsychiatry. 2002;72(2):294–302. doi: 10.1037/0002-9432.72.2.294. [DOI] [PubMed] [Google Scholar]
  49. Voisin D. Victims of community violence and HIV sexual risk behaviors among urban African American adolescent males. Journal of HIV/AIDS Education and Prevention for Adolescents and Children. 2003;5(3):87–110. [Google Scholar]
  50. Voisin D, Salazar L, Crosby R, DiClemente R, Yarber W, Staples-Horne M. The association between gang involvement and sexual behaviors among detained adolescent males. Sexually Transmitted Infections. 2004;80:440–442. doi: 10.1136/sti.2004.010926. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Voisin D. The relationship between violence exposures and HIV sexual risk behaviors: Does gender matter? American Journal of Orthopsychiatry. 2005;75(4) doi: 10.1037/0002-9432.75.4.497. [DOI] [PubMed] [Google Scholar]
  52. Walker C, Mason C. Ethnic Differences in the Effect of Parenting on Gang Involvement and Gang Delinquency: A Longitudinal, Hierarchical Linear Modeling Perspective. Child Development. 2001;72(6):1814–1831. doi: 10.1111/1467-8624.00380. [DOI] [PubMed] [Google Scholar]
  53. Wingood GM, DiClemente RJ, Crosby R, Harrington K, Davies S, Hook EW., III. Gang involvement and the health of African American female adolescents. Pediatrics. 2002;110(5) doi: 10.1542/peds.110.5.e57. [DOI] [PubMed] [Google Scholar]
  54. World Health Organization (WHO) World report on violence and health. WHO; Geneva, Switzerland: 2002. [Google Scholar]
  55. Zimet GD, Dahlem NW, Zimet SG, Farley GK. The Multidimensional Scale of Perceived Social Support. Journal of Personality Assessment. 1988;52(1):30–41. doi: 10.1080/00223891.1990.9674095. [DOI] [PubMed] [Google Scholar]

RESOURCES