Victimization and Youth Activity

by Owen Richard Carrick

decorative image: graffiti and a barred window.

This study examines the relationship between victimization and gang involvement. I used the G.R.E.A.T. dataset, which provided 5,935 students for analysis. The chosen variables analyzed gang involvement and victimization, and they controlled for the reasons why members joined gangs. I analyzed all data points using cross-tabs and chi-square tests and showed highly significant results. With 99% confidence, these results showed that gang members were 34.1% more likely to report being attacked than those not in gangs.  This, along with several additionally significant findings, led to my rejecting the null hypothesis for this study; indeed, they showed that victimization and gang involvement were correlated. Additionally, this study calls for further research into these variables and for further analysis of causality association.

Keywords: youth, gang, victimization, and protection

Violence is a large part of youth gang activity. Youth gang members engage in more violence than their non-gang peers, either with rival gang members or in victimizing others (O’Brien et al. 2013). Illegal and violent activities are key to gang activity and, along with the victimization of others, are a big part of why gangs are such a problem.  Gang initiations, codes, and members’ desire for respect can lead to violence and victimization. However, one might also argue that youth gang members are so violent because they too were once victims of violence. Joining a gang could have been their way of receiving protection from such violence. These former victims may become extremely violent in gangs because of their longing for security. That is, by providing protection, gangs give victims the means to turn their vulnerability into power, thereby turning them into the very thing that harmed them in the first place.  This cycle pushes more youth into the protection of gangs while preparing them to be the next line of violent transgressors.

To help stop the onset of youth gang involvement, the victimization of gang members must be analyzed. This analysis can help to show if victimization is a predecessor to joining a gang; if so, effective steps can be taken to intercept victims before they turn to gangs as well as to intercept violent offenders before they reach victims. This crucial period between victimization and gang involvement could put a stop both to future violence and to other illegal activities associated with gang life. With these theories proposed, this study aims to answer the question of how a person’s victimization makes him or her likely to become involved with a gang.

decorative image: boots facing each other

Review of Previous Literature

Why Youth Join Gangs

Numerous articles have attempted to explain why young people join gangs and participate in gang activities (Lenzi et al., O’Brien et al., and Taylor). However, they reached no consensus on the defining attribute or combination of attributes that led youths to gang involvement. To start, O’Brien, Daffern, Chu, and Thomas (2013) highlighted “pushes” and “pulls” that bring young people to gangs. The pushes are direct negative influences on youth behavior that serve to push them to gangs. For example, a push could be a desire for protection, the filling in for the lack of family support, or even the influence of family or friends to join a gang. The pulls are more passive benefits of gangs such as the ideas of potential monetary gain or the reputation of being a gang member. O’Brien et al. state that the pushes and pulls are equally motivating and that a combination of the two influence a young person’s decision to join a gang.

Lenzi, Sharkey, Vieno, Mayworm, Dougherty, and Nylund-Gibson (2015) divide factors for gang involvement into five, rather than two, domains: individual, family, peer, school, and community. To better fit their research measures, Lenzi et al. chose to reduce the domains by excluding community; however, they then chose to expand each domain into four main measures: empathy, parental support, peer deviance, and perceived lack of school safety. This allowed Lenzi et al. to study the more specific causes of gang involvement. Their study showed statistical significance for all four measures, with peer deviance being the biggest predictor, followed closely by feeling unsafe at school. Lenzi et al. further showed the plethora of reasons why young people join gangs.

Additionally, Panfil (2014) showed that specific behaviors can serve to amplify the factors that lead youth to gangs. For example, Panfil explored how homophobic bullying affected gang factors for youth, ultimately showing that homophobic bullying impacted reputation the most. The factor of reputation, corresponding with the individual domain mentioned by Lenzi et al., in combination with homophobic bullying led the most to fighting, violence, and eventual gang involvement. However, as with O’Brien et al. and Lenzi et al., Panfil found that gang affiliation could not be reduced to one factor. Several of the participants exhibited a combination of factors across differing domains (for example, the individual factor of reputation combined with lack of support in the family domain). Panfil ended his findings with the results that, although these factors all contributed to gang affiliation among his participants, the individual and peer factors were the most powerful. The ideas of fitting in or defending reputation were cited as strong indicators of gang affiliation. However, these indicators may not be generalizable across all populations because of the specific homosexual factor association. Nevertheless, this study generally reveals that each broad domain can be amplified or minimized by specific individual factors.

Who Gets Victimized and How It Affects Those Involved

To understand how victimization impacts gang involvement, it is important to understand why victimization happens in the first place and how it affects an individual. Kerig, Chaplo, Bennett, and Modrowski (2016) delved into this question and examined how victimization occurs, its impact on those being victimized, and its impact on those conducting violence against others. The goal of their study was to examine how being a gang member effects trauma and perpetration-induced trauma (PT) compared to non-gang counterparts. To measure this, Kerig et al. interviewed juvenile detention participants and used several unique testing measures for each of their variables. Their analysis on the victimization of gang members showed that they were more likely to have experienced and witnessed community violence when compared to their non-gang peers but that they were not more likely to meet criteria for Posttraumatic Stress Disorder (PTSD). In addition, Kerig et al. examined how gender impacted their factors and found that girls in a gang were significantly more likely to meet PTSD criteria than girls not in a gang.

decorative image: female with head bent, grayscale

One can further analyze how youth perceive the violence against them. Yonas, O’Campo, and Burke (2005) chose to interview both young boys and girls in a group setting to analyze the way they perceived the violence around them, its causes, and how their gender might account for differences. Their findings showed that the main reasons for violence were relationships, respect, gang reasons, and witnessing violence at home. Furthermore, Yonas et al. analyzed the different reasons for violence based on gender. For boys, the reasons revolved around drug usage and corresponding behavioral effects. For girls, the reasons revolved around their reputation (e.g., gossip). All these factors revealed some of the reasons violence occurs and how violence differs between people. Lastly, Yonas et. al. addressed the idea that violence breeds violence, thus supporting the victimization-gang relation.

Direct Victimization and Gang Membership 

Recent literature also examined how the roles of gang involvement and victimization affect one another. Apel and Burrow (2011) suggested that victimization is a cause for future offending, such as gang membership or carrying a handgun; such actions constitute an urge to stop that victim from being a victim again. Additionally, Apel and Burrow suggested a self-help theory whereby a person who is undertaking a negative action will do what is necessary to help himself or herself get away from that situation. With this theory, Apel and Burrow tested their data and showed several related factors. For instance, repeated bullying significantly predicted self-help factors such as gang involvement and aggravated assault but not handgun carrying. Another factor, neighborhood gunshot, was also a predictor for gang involvement and handgun carrying but not aggravated assault. Other factors, such as friends in a gang and unsupervised dating, were significant for gang affiliation and aggravated assault.

Finally, two articles provide valuable insights, especially when examined together. Taylor (2008) cited a study reporting that 57% of gang youth joined a gang for protection immediately after entering high school. Moreover, Taylor, Freng, Esbensen, and Peterson (2008) showed that gang members are at an increased risk of serious violence but experienced a slight decrease in their overall victimization. Overall, the entirety of the literature aims to explain why youth join gangs, who gets victimized, and how victimization and gang affiliation relate. Considering the preceding, how does this study further examine the victimization and gang affiliation relationship?

Current Study

In contrast to much of the preceding literature, this study examines the correlation between being a victim and becoming a gang member. While other articles have touched briefly on this topic, this study differs in its use of the G.R.E.A.T. dataset and its analysis of the direct victimization and gang relationship. If this data and corresponding information can reveal a link, then it can provide further evidence that something must be done to cut the bridge between someone suffering a violent attack and joining a gang. With these facts, the purpose of this study is to show that the victimization of a person does affect their likelihood of gang involvement.

Null Hypothesis

H0= There is no relationship between victimization and gang involvement.


H1= There is a relationship between victimization and gang involvement, in that more victimization leads to higher gang involvement.



The data for this study came from the Gang Resistance Education and Training (G.R.E.A.T.) program dataset, gathered between 1995 and 1999. This dataset was used to test the effectiveness of the G.R.E.A.T. program on the seventh-grade students who participated and was collected one year after the completion of the program when the students were in eighth grade. To determine which sites would be selected for the G.R.E.A.T. program, several factors were assessed, such as total population, population details, and location. Several sites were eliminated based on lack of implementation in select cities, lack of desired number of students, or lack of control/comparison group. These factors all limited the number of eligible choices down to 11 sites, and the questionnaires were distributed and to be administered on a specified date. This allowed for every eighth-grade student in attendance to take the questionnaire and to provide valuable data to the G.R.E.A.T. program. In total, 5,935 students from 42 different schools completed the questionnaire.

 I chose to use this dataset in this study because it provides broad aggregate data from several different locations with several schools in each location. Although this dataset did not rely on a simple random sampling, it still relied on an availability sampling based on the selected schools’ student attendance on specific days. This study then gained access to the data collected and was able to analyze every question and the answers student gave. This allowed for clear interpretation of data, which I further analyzed using SPSS (a software package used for statistical analysis) to ensure pertinent data was used for the purposes of this study.


Victimization was the independent variable for this study. I measured this by variable 185 from the G.R.E.A.T. dataset (“have you been attacked by someone with a weapon or by someone trying to seriously hurt or kill you?”), the possible answers being “yes” or “no.” The total sample for this variable was 5,935 with a total valid sample of 5,813. Of that valid sample, 5,114 (88%) answered “no” while 699 (12%) answered “yes.”

Gang participation was the dependent variable for this study. I measured this by G.R.E.A.T. dataset variable 211 (“are you now in a gang?”), the possible answers being “yes” or “no.” This gave a total sample of 5,935, with a valid sample of 5,748. Within that valid sample, 5,226 responded “no,” and 522 responded “yes.”

Lastly, I controlled for the idea of joining a gang for protection. I did so by utilizing the G.R.E.A.T. dataset question asking, “why did you join the gang?” This question gave many answers, including an option to indicate “not currently in a gang,” but for this study I only used the answer choice “for protection,” variable 229. Due to the survey’s coding, this answer was its own separate variable and thus showed whether the survey taker joined a gang for protection.

Analytic Strategy 

This study used cross-tabulations and chi-squares for analysis, because of its nominal independent and dependent variables. Cross tabulation serves to compare the independent variable to the dependent using a grid of the four possible answers, which are yes or no to the independent variable (v185) and yes or no to the dependent variable (v211). Lastly, the chi square test compares the two variables and determines how these variables fit the expected data. These two analyses, conducted by using the SPSS program, form the basis of this study’s findings.


Descriptive Statistics

In Table 1 below, the descriptive statistics, including minimum, maximum, mean, and standard deviation, are shown of the three original and three recoded variables of this study. The recode was done to make the statistics clearer. Upon the recode, Table 1 shows that the minimum for “been attacked” was 0, and the maximum was 1. This gave a mean of .12 or 12% with a standard deviation of .32528. This shows that, out of all the answers, it was more likely to have answered no to having been attacked. Additionally, for the “in gang” recode, Table 1 shows the same minimum and maximum as the “been attacked” recode, 0 and 1, but with a mean of .09 or 9%. The standard deviation of the “in gang” recode variable was .28737, and it had 5,748 total answers. This shows that respondents were more likely to have answered no to the “in a gang now” question. Lastly, the “join for protection” recode shows large differences from the original variable, but this is due to the “not in a gang” answers being removed as they did not relate to this study. This decreased the total answers from 5,853 to 723, with a minimum of 0 and a maximum of 1 instead of 2. This recode was done to show, out of those who were in a gang, how many joined for protection. This gave us a mean of .5131 or 51% and a standard deviation of .50017. These statistics show that out of all those that had joined a gang, around half of them did so for protection. So, out of our recoded total number, it was a little more likely to join a gang for protection than it was for other reasons.

Table 1: Descriptive Statistics.

Crosstabulation Results and Significance

Following the descriptive statistics, this study compared the independent and dependent variables using crosstabulation and chi-squares. Table 2 shows the results of the crosstabulation, with the most valuable data coming from comparing the percentages.

Table 2: Ever Been Attacked and In Gang Recode Crosstabulation.

Table 2 shows that, of those who answered no to “ever been attacked,” only 5.8% said yes to “in a gang now. However, out of those that answered yes to “ever been attacked,” the percentage that said yes to “in a gang” rose to 32.5%. One can also see this increase when comparing the percentages regarding the “in a gang now” variable. For all of those who answered no to “in a gang now,” only 8.8% also reported being attacked, whereas 42.9% of those who answered yes to “in a gang now” also answered yes to “ever been attacked.” This 34.1% difference is also a staggering change, and, as Table 3 shows, these results are statistically significant (X2 = 518.043, p<.001) with 99.9% confidence. This means that there is a significant relationship between victimization and gang involvement, and, therefore, the null hypothesis should be confidently rejected.

Table 3: Chi-Square Tests

Crosstabulation Results and Significance with Control Variable

I again used crosstabulation and chi-squares to find the significance of these variables with the control variable (“joined for protection”) involved. Table 4 shows the results of the crosstabulation and provides several interesting data comparisons.

Ever Been Attacked Recode and In Gang Recode and join for Protection Recode Crosstabulation

The most interesting of these points comes from the comparisons between those that answered yes or no to the question “joined for protection.” The sample that answered no to “join for protection” and yes to “been attacked” represented 38.8% of all those that answered yes to “in a gang now.” However, those that answered yes to both “join for protection” and “been attacked” represented 50.8% of those that answered yes to “in a gang now.” Additionally, there was a 17.2% increase when comparing those who answered yes to “join for protection” and “in a gang” but no to “been attacked” (63.3%) to those who answered yes to “been attacked” (80.5%). This is interesting, as there is a similar interest for those who answered no for “joining for protection”; however, the percentage increase is only 11.2%. These findings are statistically significant both for those who answered no for “joining for protection” (X2 = 4.116, p<.05) and for those who answered yes (X2 = 12.665, p<.001), but the confidence for yes is much greater and thus shows a better connection. This finding is interesting because answering no to “join for protection” shows a slightly lesser increase in percentage but still holds statistical significance. This promotes further investigation into other factors that could be leading to gang activity for youth and how those factors could relate to victimization. These findings show that there is no one cause for youth gang involvement; nevertheless, this study has found a significant relationship between victimization and gang involvement. In their totality, the statistics in the sections with and without the control variable show a 99% confidence that the null hypothesis can be rejected for this study.


Outcomes and Summary

The previous literature suggested the potential for this study’s correlation, though it was difficult to predict such a significant finding. The literature showed a correlation between gang involvement and several factors, including victimization, but the literature tended to focus on the violence that gang members experienced. This propensity towards violence for gang members, compared to their non-gang counterparts, made this current study interesting, as it shows that maybe the youth victimized in gangs were victimized beforehand and, furthermore, that this violence is a continuation of previous behaviors. This study shows that, out of all these factors, victimization is a significant contributor and needs to be taken seriously for prevention purposes. However, the determination of which factors impact gang activity the greatest should be a subject of future research. In other words, this concept necessitates further research to examine whether gang involvement is the catalyst for victimization or if victimization could be the catalyst for gang involvement.


Throughout its course, this study faced a few limitations, though they ultimately provide the possibility for growth and confidence on this topic and are not significant enough to call the study’s findings into question. The first of these limitations came from the wording of the questions used in the G.R.E.A.T. dataset. The questions in the survey did not specify when the young respondent joined a gang and, therefore, could not give a reference point to compare to their victimization. This means that the survey could not determine a timeline for when the victimization happened versus when the youth joined a gang. However, this limitation was not strong enough to counteract the statistical significance found in this study; moreover, the control variable partially controlled for it. The control variable specifically asked whether the youth joined a gang for protection or not and thus sheds some light onto the victimization and gang involvement timeline.

Additionally, the relative age of the G.R.E.A.T. dataset poses a slight limitation, especially since it is a secondary resource for this study. To further strengthen the results of this study, a current dataset or original survey could be used, preferably one focused more on victimization and gang involvement. A recent survey or dataset that asked more questions regarding victimization and when gang involvement started would be ideal. However, this study and the G.R.E.A.T. dataset are more than sufficient, and the significance of the statistics is still generalizable to today. This study’s limitation is only minor and does not require a necessary fix but instead would only be a change done to bolster these significant results.

Going Forward

This study has laid out many avenues for future research into and analysis of how victimization leads to gang membership. Analysis of causal factors, or which of these two variables is the main cause for the other, is a main area for future analysis. This is an interesting sub-focus, as it would show where the focus should be for prevention purposes. The goal of this research should be to help to reduce gangs’ negative impact, and that can be done through preventing the main cause of these negative behaviors. This is important, because prevention works best when it can impact the root cause of issues; if the root cause is victimization and not gang involvement, this highlights major implications for prevention. Furthermore, research should be conducted by comparing the various known causes of gang involvement to determine which factor or factors contribute the most. This would be beneficial for future studies to improve the prevention and treatment programs associated with them. In short, this study provides a strong basis from which to build future research.

 decorative image: concrete foundation

Going forward, a new study can be conducted using a new or recent survey to test this relation.  This new study could include the other points of future analysis and include more time identification for when gangs were joined and when victimization happened. It might also include additional factors that could influence gang involvement, test their causal relationship, and compare them individually. This avenue for future research is all quite interesting and offers only a hint of many future studies.


Apel, R., & Burrow, J. D. (2011). Adolescent Victimization and Violent Self-Help. Youth Violence & Juvenile Justice9(2), 112-133. doi:10.1177/1541204010376939.

Kerig, P. K., Chaplo, S. D., Bennett, D. C., & Modrowski, C. A. (2016). Harm as harm. Criminal Justice & Behavior43(5), 635-652. doi:10.1177/0093854815607307.

Lenzi, M., Sharkey, J., Vieno, A., Mayworm, A., Dougherty, D., & Nylund-Gibson, K. (2015). Adolescent gang involvement: The role of individual, family, peer, and school factors in a multilevel perspective. Aggressive Behavior41(4), 386-397. doi:10.1002/ab.21562.

O’Brien, K., Daffern, M., Chu, C. M., & Thomas, S. D. (2013). Youth gang affiliation, violence, and criminal activities: A review of motivational, risk, and protective factors. Aggression & Violent Behavior18(4), 417-425. doi:10.1016/j.avb.2013.05.001.

Panfil, V. R. (2014). Gay gang- and crime-involved men’s experiences with homophobic bullying and harassment in schools. Journal Of Crime & Justice37(1), 79-103. doi:10.1080/0735648X.2013.830395.

Taylor, T. J. (2008). The boulevard ain’t safe for your kids … Youth gang membership and violent victimization. Journal Of Contemporary Criminal Justice24(2), 125-136. doi:10.1177/1043986208315476.

Taylor, T. J., Freng, A., Esbensen, F., & Peterson, D. (2008). Youth Gang Membership and Serious Violent Victimization: The Importance of Lifestyles and Routine Activities. Journal Of Interpersonal Violence23(10), 1441-1464.

Yonas, M. A., O’Campo, P., & Burke, J. G. (2005). Urban youth violence: Do definitions and reasons for violence vary by gender? Journal Of Urban Health82(4), 543-551.