Article Text

## other Versions

Risk and protective factors associated with gang affiliation among high-risk youth: a public health approach
1. Dawn Delfin McDaniel
1. Correspondence to Dr Dawn D McDaniel, Epidemic Intelligence Service Officer assigned to the Division of Violence Prevention, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, 4770 Buford Hwy NE, MS F64, Atlanta, GA 30341, USA; dawn.mcdaniel{at}cdc.hhs.gov

## Abstract

Background Gang violence accounted for 20% of homicides in large cities from 2002 to 2006. Preventing gang affiliation (ie, youth who either desire or have gang membership) might reduce subsequent gang activity. Previous research has focused on identifying risk factors for gang affiliation; however, little information is available on protective factors.

Aim To identify risk and protective factors to provide more direction for gang violence prevention strategies.

Methods The author analysed cross-sectional survey data from 4131 youths in grades 7, 9, 11 and 12. Data were collected in 2004 from students in a high-risk, urban public school district. Regression analyses were conducted to assess the association between gang affiliation and alcohol and drug use, delinquency, depressed mood, suicidal ideation, peer victimisation, parental monitoring and positive reinforcement, adult, family and peer support, coping skills, and school connectedness. Analyses were controlled for sex, race/ethnicity and age.

Results An estimated 7% of youths were gang affiliated. Adjusting for all factors, gang affiliation was positively associated with engaging in any delinquent behaviours (prevalence OR: 2.07; 95% CI 1.18 to 3.64), frequent alcohol use (OR: 2.62; 95% CI 1.85 to 3.72) and frequent drug use (OR: 1.95; 95% CI 1.15 to 3.29). Gang affiliation was negatively associated with moderate levels of parental monitoring (OR: 0.67; 95% CI 0.54 to 0.85) and coping skills (OR: 0.54; 95% CI 0.42 to 0.71).

Conclusions The findings suggest the potential benefit of increasing parental monitoring and coping skills and reducing delinquency, alcohol use and drug use to prevent gang affiliation.

• Gang
• protective factors
• risk factors
• primary prevention
• youth
• street youth
• public health
• violence
• risk/determinants

This is an open-access article distributed under the terms of the Creative Commons Attribution Non-commercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license. See: http://creativecommons.org/licenses/by-nc/2.0/ and http://creativecommons.org/licenses/by-nc/2.0/legalcode.

## Introduction

From 2002 to 2006, gangs were responsible for approximately 20% of homicides in the 88 largest US cities,1 and gang violence has been perceived as reaching ‘epidemic’ levels in many urban areas.2 Furthermore, gang members are more likely than their non-gang-affiliated peers to engage in crime and violence,3 4 which increases their risk of violence-related injuries and death.5 6 While only an estimated 5% of the US population has ever joined a gang,7 gang membership has reached 14–30% of the population in many urban areas.8 9 The impact of gang membership on a youth's health is tremendous and preventing gang membership or the desire for gang affiliation early in adolescence might significantly reduce subsequent gang-related violence. This study explores a public health approach to understanding the factors that increase risk for, as well as protect against, gang affiliation among youth who reside in a high-risk community (ie, a community with high levels of serious crime, poverty and unemployment).

Traditionally, initiatives designed to reduce gang affiliation have not focused on primary prevention strategies; that is, strategies aimed at preventing gang affiliation before gangs begin recruiting youth. These strategies often attempt to prevent factors that might increase risk of gang affiliation while promoting factors that might lower risk. Risk factors for gang affiliation include friendships with delinquent peers, lack of parental monitoring, negative life events (eg, loss of parent), and positive attitudes towards delinquency or engaging in delinquency (eg, violence perpetration).7 10 Other risky behaviours associated with gang affiliation include alcohol use and drug use.11 To date, little research has focused on mental health related factors, such as depression and suicidal ideation;11 these factors might be important since they are common among youth who engage in delinquency.12 Several studies have also shown that the probability of gang affiliation increases as risk factors accumulate.10 11

In contrast to the amount of literature on risk factors for gang affiliation, the literature on protective factors is limited. (The term ‘protective factor’ often refers to a factor that buffers the effects of risk. However, for this study, the term refers to factors found to be negatively associated with gang affiliation.) The few studies that have embarked on this topic have, in general, found that youth who have strong social skills, a supportive family (eg, parental monitoring, warmth and control) and connections with religion/religious institutions are less likely to seek gang affiliation than their peers13–15; however, more research is still needed to understand the protective factors for youth in high-risk communities, especially with regard to understudied factors, such as positive peer support and school commitment.16 This study assesses risk and protective factors associated with gang affiliation among a high-risk youth population to better inform primary prevention strategies.

## Methods

### Study design

This study used data provided by the Centers for Disease Control and Prevention's ‘Youth Violence Survey: Linkages among Different Forms of Violence’ study.17 The Youth Violence Survey was a cross-sectional study of youth in a high-risk community that examined various types of violence and their common risk and protective factors. These factors addressed the different domains of the social–ecological model, which includes individual, peer, family and community level factors.18 This survey also captured information on gang affiliation.

### Study population

The Youth Violence Survey was administered to students in grades 7, 9, 11 and 12 in a high-risk community. The high-risk community was identified using multiple indicators of risk (ie, nationally the school district was among the highest 25 in poverty, the highest 15 in single-parent families, the highest 10 in serious crime rates and the highest 35 in rate of unemployment). This site was selected based on these demographic characteristics and pragmatic considerations (eg, commitment to study and feasibility of obtaining adequate census of students in targeted grades). While the site will not be named to protect its confidentiality, the community was racially and ethnically diverse with a population of less than 250 000. There were 16 public schools within this community's school district and all schools participated.

The study received Institutional Review Board approval from the Centers for Disease Control and Prevention and ORC Macro International. Prior to data collection, written parental consent and student assent were obtained from all students: 14% of students did not return the consent form and 1% of students refused to participate. Non-English speaking parents were offered consent forms in the major languages represented in the community, including Spanish.

Data collection for this study was undertaken in April 2004. Students voluntarily completed the anonymous self-administered 174-item questionnaire during one class period. Students received a $5 gift card for returning the parental permission form (regardless of whether parental permission was granted) and another$5 gift card after completing the survey. Those students who did not want to participate were assigned individual deskwork by their teacher. Among the 5098 students who met eligibility for the study, a total of 4131 (81%) chose to participate: 1491 students were in grade 7 (83% participation rate), 1117 students were in grade 9 (73.4% participation rate) and 1523 students were in grades 11 and 12 (79% participation rate). Further information about the study design, the population and the setting is available in the study by Swahn et al.17

### Measures

#### Outcome variable

The outcome variable for this study, gang affiliation, was based on one item in which students were asked how they felt about joining a gang. The students were not given clarification regarding the definition of a gang. Responses were dichotomised to distinguish between gang-affiliated students who reported either the desire to join a gang or have an active gang membership with no intention of leaving the gang and non-gang-affiliated students who reported no desire to join or actively participate in a gang.

#### Main explanatory variables of interest

Thirteen variables were examined to identify their association with gang affiliation. These variables spanned the domains of individual, peer, family and community level characteristics. Six factors were assessed to determine whether they had a positive association (ie, act as potential risk factors). These factors included having a depressed mood in the past 30 days, having suicidal ideation in the past year, having consumed alcohol at least once in the past year, having used drugs at least once in the past year, engaging in any delinquent behaviour in the past year and having been victimised in the past year. Seven variables were assessed to determine whether they had a negative association with the outcome variable (ie, act as potential protective factors). These variables included having the confidence to cope with conflict, having peer support, having parents who have used positive reinforcement in the past 30 days, having parents who have used monitoring strategies in the past 30 days, having family support, having adult support at school and feeling connected to school.

Details of how these variables were coded from the survey are shown in the online supplementary appendix A. In general, these variables exhibited high internal consistency (see Logan et al for further description19). Multi-item variables were re-classified to account for the distribution of the data in a conservative manner while retaining variability when possible.

### Statistical analysis

Basic descriptive analyses were conducted to characterise non-gang- and gang-affiliated students. Logistic regression was used to assess crude associations between the outcome variable and demographic variables as well as the main variables of interest. Associations were reported via prevalence ORs. Robust variance estimates were used to account for potential effects from cluster correlation within each school. Multiple logistic regression was used to determine the adjusted ORs for each of the main variables of interest. This model adjusted for demographic factors (ie, age, sex and race/ethnicity) to control for potential developmental and sociocultural influences.20 Since youths can have more than one risk or protective factor, we also adjusted for all of these factors in the model to provide more accurate individual associations between the factors and the outcome variable. Furthermore, we assessed for colinearity using variable inflation factor scores. Variable inflation factor scores fell within acceptable ranges (<2), which suggests that colinearity among the variables used in the models was not an issue.

The proposed risk and protective factors were summed to create cumulative indices. For these indices, all variables were rendered dichotomous (see online supplementary appendix A for more details). Multiple logistic regressions were used to determine the association of these indices with gang affiliation, controlling for age, sex and race/ethnicity. Also, the cumulative risk index was dichotomised into a risk grouping variable by a median split to distinguish those at high risk (≥3 risk factors) from those at low risk (≤2 risk factors) for subsequent analyses. This variable was used in a model that included interaction terms between this risk grouping variable and the salient protective factors from the adjusted model.

To illustrate the potential benefit of having the salient protective factors at high and low levels of risk, another multiple logistic regression was run including the risk grouping variable and key protective factors. The estimated predicted probabilities for gang affiliation were determined for youth at high and low risk with no protective factors and with each combination of protective factors. Fisher's exact tests with a Bonferroni correction for multiple comparisons were used to compare these probabilities.

## Results

Descriptive characteristics of the study population are reported in table 1. The majority of this youth population was female, of ethnic minority status and of high school age (ie, 14–18 years or older). An estimated 7.1% of the youths were gang affiliated: 2.4% of the youths reported the desire to join a gang and 4.8% of the youths reported active gang membership with no intention of leaving the gang. A majority also reported having consumed alcohol in the past 12 months (48.2%), having participated in delinquent behaviour in the past 12 months (62.2%) and having been victimised by a peer in the past 12 months (54.5%). However, a large proportion of the study population also reported having at least a moderate level of confidence to cope with conflict (64.3%), having parents who have provided positive reinforcement in the past 30 days (63%), having parents who monitored their activities at least at a moderate level in the past 30 days (64.2%), having adult support at school (93.5%) and feeling connected to school (79%).

Table 1

Frequencies and crude ORs for demographic, risk and protective factors for gang affiliation (N=4131)

The crude associations between the predictor variables and the outcome variable in table 1 revealed that gang affiliation was most common among youth of ethnic minority status, males, those who had a depressed mood, reported suicidal ideation, used alcohol or drugs, were delinquent, and had experienced peer victimisation. Gang affiliation was least common among youth who had confidence in their coping skills, peer support, parental positive reinforcement, parental monitoring, family support, adult support at school and school connectedness.

The adjusted ORs for all predictor variables are presented in table 2. Alcohol use that occurs at least 2–3 days a month, drug use that occurs at least 1–2 days a week and any delinquency remained positively associated with gang affiliation. Having at least a moderate level of confidence with regard to coping with conflict and having at least a moderate level of parental monitoring remained negatively associated with gang affiliation.

Table 2

Adjusted ORs and 95% CIs for demographic, risk and protective factors for gang affiliation (N=4131)

The findings on cumulative risk and protective factor indices are presented in table 3. Over half of the study population reported two or more risk factors. Those youths with four or more risk factors were nearly six times more likely to be gang affiliated than those with zero to one risk factor. Also, those with three or less protective factors were 5.6 times as likely to be gang affiliated than those who had four or more protective factors reported. The cumulative risk and protective factor indices were analysed in an adjusted model, controlling for demographic characteristics. The odds of gang affiliation were significantly higher as a youth accumulated more than two risk factors; however, as the youth acquired over five protective factors the odds of being affiliated with a gang was significantly lowered. No significant associations were found for the interaction terms between risk level and the significant protective factors, suggesting that these protective factors did not buffer risk. Instead, the protective factors had a similar negative association with gang affiliation for all youth regardless of risk level.

Table 3

Frequencies and adjusted ORs for cumulative risk and protective factors for gang affiliation (N=4131)

The predicted probabilities for gang affiliation based on various combinations of the significant protective factors (ie, confidence in coping skills and parental monitoring) and risk level are presented in table 4. Youth who reported having either coping skills alone or coping skills in combination with parental monitoring had a significantly lower probability for gang affiliation than those youth with neither protective factor, regardless of risk level (p<0.01). Also, youth who were considered high risk with both protective factors had significantly lower probabilities for gang affiliation than youth at low risk who did not have either protective factor (p<0.05). Youth at high risk with coping skills alone had similar probabilities for gang affiliation as youth at low risk with neither protective factor.

Table 4

Predicted probabilities of gang affiliation for youth with various combinations of the significant protective factors and risk level

## Discussion

The finding that gang affiliation is associated with engaging in delinquent activities, particularly illicit alcohol and drug use, is one of the most consistent findings in the gang literature.4 11 While the temporal relationship of these variables cannot be explained using the current study design, prior research using longitudinal data has shown that delinquency and illicit substance use have an ‘enhancing’ relationship with gang affiliation.11 For example, it has been found that early involvement in these antisocial behaviours elevates the risk for subsequent gang affiliation by reducing the possibility for positive social connections and increasing opportunities for delinquent peer associations. Furthermore, the reciprocal relationship is also thought to exist where gang affiliation ‘facilitates’ an increase in these antisocial behaviours (eg, excessive substance abuse, more severe criminal behaviour) through peer pressure.11 Reducing the risk of initiating illicit substance use or delinquency early in preadolescence or adolescence might disrupt the initial trajectory towards gang affiliation or even potentially break current gang-related connections that could further escalate these behaviours.

This study also showed the potential benefit of emphasising protective factors in prevention efforts. Two protective factors, youth's confidence in their coping skills and parental monitoring, were negatively associated with gang affiliation. Youths at high risk who had the two key protective factors had lower probabilities for gang affiliation than youths with neither of these factors at low risk. Also, youths at high risk with only one protective factor, confidence in their coping skills, had similar probabilities for gang affiliation as youths at low risk with neither factor. For youths in high-risk contexts, a focus on reducing risk factors may not be feasible in all situations due to the number of risk exposures and the difficulty in changing certain risk factors. Therefore, if it is not possible to reduce risk, adding one or two key protective factors may make youth at high risk appear similar to those at low risk. Other studies have documented a similar value in the use of protective factors against other forms of violence.21–23

### Limitations

This study was subject to several limitations. First, the cross-sectional nature of these data limited our ability to understand the causal ordering of the risk and protective factors and the outcome, gang affiliation. For instance, we were unable to determine whether confidence in coping with conflict prevented gang affiliation or whether youth developed poor coping strategies while affiliated with gangs. We also did not attempt to disentangle the complex relationships between gang affiliation, delinquency, alcohol use and drug use. Future prospective research would provide a better understanding of the relationship of these correlates. Second, surveying gang-affiliated youth in school may introduce a selection bias, and a non-response bias analysis was not done in this study. Further work that assesses whether these findings can be replicated in samples of youth who have dropped out of school, are truant or are adjudicated is important. Third, the survey did not define gang affiliation for respondents; therefore, gang affiliation in this study is likely to have a broad meaning. Past research has shown that using a more restrictive definition leads to more precision in distinguishing gangs from other youth groups.7 Finally, this study is limited to a high-risk population and the findings cannot, by this study design, be generalised to other populations.

### Prevention strategies

This study has several important implications for prevention efforts. Since youth with both parental monitoring and coping skills had the lowest probability of gang affiliation, programmes that offer parent training that are focused on skills relevant to effectively monitoring children, in addition to helping youth develop strategies to cope with conflict, may be most beneficial for preventing gang affiliation. Several primary prevention programmes aimed at reducing delinquent behaviour have shown positive effects.24 25 Examples of evidence-based prevention programmes that address coping skills (ie, interpersonal skills, such as talking out disagreements, and intrapersonal skills, such as controlling anger) include the Promoting Alternative THinking Strategies (PATHS) curriculum and I Can Problem Solve.24 25 Examples of concurrent parent and child programmes targeting both coping skills and effective parenting include Strengthening Families Program, Fast Track, Prevention Treatment Program and Incredible Years.25 Of these programmes, only the Prevention Treatment Program showed reductions in gang affiliation.26 While the other programmes have not evaluated their impact on gang affiliation, given the shared aetiology of delinquency, drug and substance use, and gang affiliation, these programmes would likely be promising.

### Future directions

Using a public health approach, we identified the potential of increasing a youth's confidence in their ability to cope with conflict and parental monitoring and reducing their delinquency, alcohol use and drug use to prevent gang affiliation. While studies that identify risk and protective factors represent one of the initial steps of the public health model, further work needs to be done to develop and evaluate primary prevention programmes that help reduce gang affiliation. Also, more work is needed to help disseminate programmes that can potentially prevent gang affiliation. Dissemination may be best accomplished with the inclusion of public health professionals in the planning and implementation of such programmes. These public health professionals may be able to assist in the necessary on-going collaboration with other sectors that traditionally work with gang-affiliated youth, such as criminal justice, to be most strategic about ways of combining community efforts to reduce gang affiliation.

### What is already known on the subject

• Risk factors for gang affiliation include friendship with delinquent peers, lack of parental monitoring, negative life events and positive attitudes towards delinquency or engaging in delinquency.

• Other risky behaviours associated with gang affiliation include alcohol use and drug use.

• The probability of gang affiliation increases as risk factors accumulate.

• This study showed the potential value of emphasising protective factors when assessing factors associated with gang affiliation.

• The findings suggest the potential benefit of increasing parental monitoring and coping skills and reducing delinquency, alcohol use and drug use to prevent gang affiliation.

## Acknowledgments

I wish to thank Dr J Logan for his contributions to this research.

View Abstract
• ## Supplementary Data

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Files in this Data Supplement:

## Footnotes

• The findings and conclusions in this manuscript are those of the author and do not necessarily represent the views of the Centers for Disease Control and Prevention/the Agency for Toxic Substances and Disease Registry.

• Funding This study was supported by Centers for Disease Control and Prevention.

• Competing interests None.

• Ethics approval Ethics approval was provided by Centers for Disease Control and Prevention (CDC) and ORC Macro International.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Data sharing statement The data were obtained from the CDC's Youth Violence Survey.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.