Volume 3, Issue 2, June 2002

ISSN 1096-4886 http://www.westerncriminology.org/Western_Criminology_Review.htm
© 2002, The Western Criminology Review. All Rights Reserved.

Variables Affecting Adolescent Victimization:
Findings from a National Youth Survey 

Gideon Fishman

Gustavo S. Mesch

Zvi Eisikovits


Citation: "Fishman, Gideon, Gustavo S. Mesch and Zvi Eisikovits, "Variables Affecting Adolescent Victimization: Findings from a National Youth Survey." Western Criminology Review 3 (2). [Online]. Available: http://www.westerncriminology.org/documents/WCR/v03n3/fishman/fishman.html



The purpose of this study Is to investigate the link between offending and victimization among an adolescent population. Data from the 1998 Israeli National Youth Survey are used to test two conflicting hypotheses. One, derived from the extensive literature on bullying behavior, suggests that there is a clear distinction between those who commit acts of aggression and those who are on the receiving end. The other, which is the more criminologically sound and draws on the subcultural approach, holds that those involved in aggressive behavior also have a higher likelihood of becoming victims of aggression.

The findings show that engaging in aggressive behavior increases the likelihood of being on the receiving end. Adolescents reporting that they had slapped or hit their friends were also more likely to be victims. The high association between victimization and aggressive conduct can be explained by the simple fact that when people are hit or insulted they are most likely to hit back in order to defend themselves. However, there is a distinct category of youth that does not report aggressive conduct, and who also seem to have a low probability of becoming victims. This might cast some doubt about the approach that tends to imply that a division of labor exists between perpetrators and victims.

The findings also provide support for the subculture of violence approach and the proximity hypothesis. Adolescents who know delinquent peers and live in neighborhoods where violence against children is common are more likely to be victims of both verbal and physical aggressive behavior.  


Keywords: victimization, bullying, National Youth Survey, delinquency, criminality

Variables Affecting Adolescent Victimization:
Findings from a National Youth Survey 

Aggressive behavior in schools and among adolescents is an extensive and serious problem that affects a relatively high proportion of students. Recent studies have identified important consequences of violent youth victimization. Youth victimization leads to feelings of distress, unhappiness, and loneliness, increased social isolation and negative attitudes toward school, and to physical injuries (Boulton and Underwood 1992).

Violence at schools and among peers in general has become a very common phenomenon in Israel and a cause for concern to educators, youth workers, law enforcement agents, and policy makers. This study attempts to identify who the victims of violent conduct are and what might account for their victimization.

What affects victimization, especially among youth, has long intrigued law enforcement agencies, educators, and policy makers, and it is the main focus of this paper. Age has been established as a central variable accounting for victimization. Surveys show that young males are at a higher risk of victimization than other age categories (Sampson and Lauritsen 1994).

Although youth are at the highest risk of victimization, juvenile victimization appears to be less harmful overall, in terms of severity of injury and monetary loss (Lauritsen, Sampson, and Laub 1991).

Despite the importance of age in explaining violent victimization, relatively few studies have explored the covariates of youth victimization. The purpose of the current study was to fill a gap in the literature and to explore the link between aggressive conduct and victimization. Specifically, the question asked was whether youth who report aggressive behaviors are also more likely to report violent victimization in the same time frame.

The current study enabled us to examine two conflicting hypotheses. One, derived from the literature on bullying behavior, suggests a clear distinction between those who commit acts of aggression and those who are on the receiving end: perpetrators and victims are, predominantly, two distinct and exclusive groups. The second hypothesis suggests that those involved in aggressive behavior have a higher likelihood of becoming victims of aggression. A complementary point should be emphasized here, namely, that those who stay away from aggression have a low likelihood of becoming victims. Whereas getting into fights may cause victimization without real regard as to who is the aggressor, if indeed those who shy away from aggressive conduct have low likelihood of becoming victims, this may put into question the notion of division of labor between victims and perpetrators. In addition, and beyond testing these hypotheses, this study allowed an exploration of the variables that increase the risk of victimization and the contribution of aggressive behavior to violent victimization among adolescents.


In the last decade the public discourse in Western societies has been very much concerned with violence at school. A current discussion of violence inflicted on pupils by each other is a forerunner of attempts to abolish this form of peer violence in schools. Although physical attacks between peers were traditionally categorized as "rough play" (Boulton 1994) and integrated parts of pupils’ culture, we have become critical of this type of children’s behavior and are moving to de-legitimize it both inside and outside school (Zinnecker 1998). There is now a tendency to overemphasize dramatic aspects of school life, such as gang delinquency and the use of weapons, within school territories. Other arguments that are popular in the context of these media campaigns and moral panics are statistics about a dramatic increase of violence among pupils, although the statistical data are not always clear or are even missing (Best 1994). Part of the academic concern with bullying is the perception that bullying and delinquency are closely related. It has been argued that bullying and delinquency may overlap because they are both indicators of same underlying theoretical construct, such as antisocial personality. In this approach, bullying may be an earlier stage of a developmental sequence leading to delinquency, but the empirical evidence is still scarce (Baldry and Farrington 2000).

Public opinion has shifted its attention in recent years from perpetrators to victims. This tendency is paralleled with the growing interest in victimization and particularly the concern for children as victims (Best 1990). In the case of adolescents, this means increased public concern with safety standards and quality of life of children and adolescents in schools, classrooms, and schoolyards, and the generation of numerous programs with the aim to reduce bullying and peer rejection.

The extensive research literature on bullying among high school students suggests that offenders and victims are two distinct populations. Studies found that bullying behavior is common among adolescents. Studies conducted in different European countries show that the prevalence of bullying is quite high: between ten and twenty percent of the adolescents reported being physically assaulted (hit or kicked) during the course of one year (Farrington 1993). The term "bullying" refers to physical, verbal, and psychological attacks intended to induce fear and distress or to inflict physical harm upon the victim (Farrington 1993). Physical violence is one of the different forms of behavior under the label of bullying (Cohn 1987).

Some of the literature on bullying suggests that the perpetrators are children frustrated by their lack of success in school. They build their reputation using aggressive behavior against other adolescents who are physically and socially weaker. The most attractive targets of aggressive behavior are females and younger adolescents who are doing well in school. The aggressors tend to be males, and relatively older adolescents from lower socio-economic status groups. According to this view, offenders and victims are thus two different groups that interact in a context of power imbalance. The more powerful attack the less powerful in a location that provides opportunities to carry out the assault, such as playgrounds and remote hallways where adults are not present (Farrington 1993). There is some empirical evidence that bullies and their victims might be two different groups. Olweus (1991) found that only one bully in ten was also a victim, while only one victim in eighteen was also a bully. Another study reported that twenty percent of the victims were also bullies, and that their aggressive behavior was always directed against children who had not offended them (Roland 1989). A recent longitudinal study reported a relatively high correlation of .44 between amount of bullying and amount of being bullied, which suggests that being a perpetrator enhances the chances of also being a victim (Zinnecker 1998). But some have argued that the overlap between offenders and victims is in reality even lower, because offenders might over-report their victimization in order to justify their acts. While seemingly logical, this hypothesis was not supported by empirical evidence (Olweus 1991).

The underlying assumption that the two groups are different rests on the power imbalance idea. This imbalance may have physical and social dimensions. In terms of physical characteristics, studies have shown that victims of bullying report low scores in the physical self-concept subscale. Victims have been found to be physically relatively weak (Salmivalli 1998). One indicator of physical strength is age: older adolescents tend to be taller and stronger (Olweus 1991). Therefore, the younger the adolescent, the higher his/her likelihood of belonging to the victim category.

Another indicator of physical strength is gender. Boys are likely to be more aggressive than girls, especially in displaying direct forms of aggression such as hitting, whereas girls are more likely to become victims (Baldry and Farrington 1998).

The literature on bullying portrays the victims and offenders as very different in terms of popularity and peer acceptance. Perpetrators report high scores on social and physical self-concept. Bullies seem to think they are physically rather fit, and they also view themselves as being quite popular among their peer group. Victims report low scores in two self-concept sub-domains, namely the social and physical (Salmivalli 1998). They feel unpopular and rejected by their peers. They lack social skills, and their relationships with friends are weak. Adolescents who lack good friends have no support when exposed to an offender's aggressive behavior (Farrington 1993).

Studies have shown that the family plays a central role in the development of social skills in youth. Parents who are supportive and involved with their children enable them to develop personal and social skills (Rigby 1996). Although there has been little research examining adolescent-parent relationships and victimization, a number of studies have shown that adolescents who have authoritarian and non-supportive parents are more likely to be victims of aggressive behavior (Baldry and Farrington, 1998; Bosworth, Espelage, and Simon 1999)

In sum, the literature on bullying suggests that an overlap between offenders and victims is very unlikely. Also, victims are likely to be young and female, to report distant and non-supportive relations with their parents, to have very few close friends, and to express positive attitudes toward their teachers and school.

The victimization literature, on the other hand, presents different results. Studies on youth violent victimization, contrary to the power imbalance hypothesis, have shown an overlap between violent offending and violent victimization (Esbensen and Huizinga 1991; Wells and Rankin 1995). A comprehensive paper based on the analysis of the 1976 National Youth Survey (NYS) and the 1987 Monitoring the Future (MTF) data came up with very different results. According to the NYS, in the US only twelve percent of the non-delinquent youth surveyed reported victimization by assault, while forty-five percent of the delinquent youth reported such victimization. Accordingly, offenders were more than three times as likely to be assaulted than were non-offenders. This pattern is also shown in the MTF study, where offenders were more than twice as likely to be assaulted than non-offenders (Lauritsen, Laub and Sampson 1992).

One possible explanation is grounded in the lifestyle-routine activity theory. According to this approach, certain demographic characteristics increase the risk of victimization because role expectations are related to a lifestyle that places suitable targets in proximity to motivated offenders without appropriate adult supervision (Meier and Miethe 1993). Based on this argument Lauritsen, Sampson, and Laub (1991) suggest the principle of homogamy. Namely, persons are at a higher risk of victimization when they come into contact with members of demographic groups that contain a disproportionate numbers of offenders. By this logic, males and older adolescents are more likely to be victims than females and younger adolescents (Finkelhor and Asdigian 1996). Furthermore, as expected, older adolescents, more often than other age groups, pursue activities outside their homes and are more likely to spend time at parties, bars, and other places where young people congregate routinely. Therefore their risk of victimization is higher (Lauritsen et al. 1991). Support for this hypothesis is presented in a study that found a positive association between activities such as going to parties and bars and violent victimization among high school students (Jensen and Brownfield 1986). Participating in nighttime social activities outside the home brings individuals into proximity with potential offenders. Such activities are often unsupervised, and guardianship that might be provided by parents or other responsible adults is lacking.

Much of the research on youth victimization has emphasized the connection between delinquent activities and victimization. Delinquents have been found to be more prone to victimization than non-delinquent youth. Individuals who engage in deviant activities are at a higher risk of becoming victims of a personal violent crime, and the longer the time spent with delinquent peers the higher the risk of victimization (Lauritsen et al. 1991). Delinquency is perceived as a lifestyle since it is not a sporadic event but a continuous activity. It takes place in unconventional social settings where other delinquents congregate and where supervising adults are not present.

The evidence that violent conduct and a deviant lifestyle increase the risk of personal victimization seems convincing. A multivariate analysis showed that mildly violent offending is directly related to the risk of victimization (Sampson and Lauritsen 1994). Another study found that adolescents who reported involvement in delinquency were two to three times more likely to be victims of assault than youth who reported no delinquent involvement. Moreover, having delinquent peers had a highly positive effect on victimization by a personal violent crime, even when variables such as race, gender, family structure, family income, and neighborhood characteristics were controlled for (Lauritsen et al. 1991). On the other hand, adolescents who spent more time with their family were less likely to become victims of aggressive behavior. Similarly, adolescents who reported satisfaction with school activities were less likely to report violent victimization (Lauritsen et al. 1992).

The literature tells us that "good kids" stay out of trouble, while violent offenders seem to make good targets for victimization. Often such victimization goes by with relative impunity. Offenders are more vulnerable to victimization because they are less likely than victims who are not offenders to report to the police (Sparks 1982).

Findings from the lifestyle routine activities approach indicate that the perpetrators and the victims might not constitute exclusive groups, and there is an overlap between violent offending and violent victimization. In addition, certain adolescent activities and characteristics seem related to victimization risk. While age, gender, and having delinquent peers increase the risk of violent victimization, conventional activities such as spending time with parents, attending school, and holding positive attitudes toward school reduce that risk.

The theory of a subculture of violence also points to the overlap between violent offending and victimization (Wolfgang and Ferracuti 1982). According to this approach, for certain areas and population groups, a system of subcultural values exists that supports the use of violence. This value system condones violent response in a situation where normal culture would not allow it. Singer (1981) found that where a subculture of violence is prevalent there is a high likelihood that an individual will alternate between the role of offender and victim. By this approach, violent offenders and victims of violent victimization are related to and resemble each other because both live in an environment where violence is common and accepted. Such findings also fit the "proximity" hypothesis (Fagan, Piper, and Yu-The Cheng 1987), which argues that the correlation between offending and victimization is high due to the proximity of residence of the victims and perpetrators. The proximity hypothesis claims that independently of individual characteristics, neighborhood violence represents an important determinant of risk (Lauritsen et al. 1991).

To sum up the review, two sets of theoretical approaches seem to account for violent victimization of adolescents by their peers. One explanation, which has dominated the literature on bullying, argues that it is a result of frustration of youth due to their failure in school, and that because of power inequality, these youth prey on other adolescents who are weaker and less assertive. This argument is supported by research in schools on normal populations. However, another significant body of literature, derived from lifestyle-routine activity approach and also from the subculture of violence approach, indicates that the most likely victims are delinquents or youth who, due to lifestyle, spend much time in proximity to other delinquents or violent youth. According to research findings, there is an overlap between perpetrators and victims. This study attempted to examine which hypothesis is most suitable for the Israeli situation when applied to normal, high school-age adolescents.


This study was a part of an annual national youth survey conducted by the Minerva Center for Youth Research at the University of Haifa, Israel, based on a representative sample of 1000 households in this country. The sampling procedure began with a random sample of sixty out of all localities in Israel with a population of 2000 residents or more. Then, according to the size of the adolescent population in each locality, neighborhoods were selected randomly. The number of neighborhoods in each selected locality was determined by the size of the juvenile population (13-18 years old) there. At least one neighborhood was randomly selected in places where there was a low rate of adolescents, and more than one in the larger urban areas. In each neighborhood, fifteen households were randomly selected. The selected neighborhoods represent all geographic areas of Israel, and also different sizes of localities, from big cities to small towns and villages.

Data Collection

The designed questionnaire was not pre-tested this time since the previous annual survey was taken as a test for the validity of the questions. Questions that had seemed unclear or required clarification were dropped or modified. The data were collected in 1998. Trained university students who contacted each randomly selected household conducted the survey. If no youngster aged 13-18 was in the household, the interviewer replaced the household by another in the same building or in the building across the street. Apartment buildings and streets are known to be highly homogeneous in terms of socio-economic status composition, family status and ethnicity; therefore it was reasonable to expect that no biases would result. Face-to-face interviews were conducted, in which a structured questionnaire was used. The interviewee was guaranteed total anonymity, and the interview lasted about 40 minutes. Overall, 1,000 youngster were contacted; 114 refused to participate, representing a refusal rate of 11.4 percent. The original sample included Israeli Arabs and Jews, but the number of Arab adolescents was too small to conduct separate analysis. Therefore the sub-sample of Israeli Jews only was used (n=740).

Two dependent variables were examined. One was violent victimization, which refers to violence inflicted by peers either at school or in the victim's neighborhood. Youngsters were asked whether in the previous year they had been hit, slapped, or kicked either at school or in their neighborhood by their peers. The possible response was either yes or no. The second dependent variable referred to verbal abuse. The question posed here was whether peers had insulted, cursed, or said unpleasant or nasty things to them, either at school or in their neighborhood during the previous year. Here too the response was either yes or no.

The independent variables were:

  1. Family relations: this was a composite variable of three questions dealing with how well parents showed interest in and listened to an adolescent’s problems, how close adolescents felt to the family, and to what extent the respondents felt that they got along well with their parents. All questions were placed on a Likert scale and were factor-analyzed. The three items were scored on one single factor, and a composite scale was created. The internal reliability of the composite scale was .815.
  2. Relations with peers: this too was a composite variable. They responded to three questions: how important was it for the respondent to spend time with friends; had the respondent had close friends for a long time; and were the friends willing to listen to the respondent’s problems. The responses were measured on a Likert scale and were factor-analyzed. A composite scale was created with internal reliability of .646.
  3. Peers’ delinquency: this was a composite variable that asked respondents if any friends they knew engaged in behavior such as "breaking-into and stealing from cars," "breaking-into apartments," and "shoplifting." Factor analysis showed that the items scored on a single factor. The items were used to create a composite measure with an internal reliability of .790.
  4. School satisfaction: this was a composite scale created out of responses to three items that sought the level of agreement of the respondent with three statements: "My teachers present a good role models," "I respect most of my teachers," and "My classes at school are interesting." The responses were measured on a Likert scale and resulted in a single dimension when subjected to factor analysis. The items were combined into a single measure with a reliability coefficient of .776.
  5. Perception of violence as common in the neighborhood: this was measured by a direct question about the prevalence of parents in the neighborhood using violence against their children (beating them).
  6. Involvement in aggressive behavior: two questions were used to measure involvement in aggressive behavior. One asked the respondent if in the previous year he/she had cursed, insulted, or said nasty and unpleasant things to other friends. The other item asked if in the previous year the respondent had slapped or hit one of his/her friends. A ‘yes’ response was coded 1 and a ‘no’ response was coded 0.

Because the dependent variables of the study were measured as dummy variables, a logistic regression was used in the multivariate analysis.


The main characteristics of the sample population are presented in Table 1. The average age of the respondents was 16.29 years. Close to 53 percent were females and 47 percent were males. The use of aggressive verbal language appeared to be quite common among adolescents. Almost 44 percent of the sample reported that during the previous year they had been verbally victimized; namely, peers had insulted them and said offensive things to them. Victimization by physical violence was less common, as 13 percent of the respondents reported having been slapped, hit, or kicked by a peer during the previous year. As to the bullying behavior of the respondents themselves, 27.3 percent reported that they have insulted or said nasty things to their peers, and 9.6 percent reported that they had hit or slapped a peer during the previous year. However, from the descriptive results it is evident that a higher proportion of the sample reported being victimized by others, physically and verbally, than victimizing others (see Table 1).


Table 1 

Characteristics of the Sample

Mean age

Std. dev.










Bullying - Victimization (in percent)

Did your peers insult or say nasty and unpleasant things to you in the last year?




Did any of your peers slap or hit you in the last year?


Bullying - Offending

Did you slap or hit one of your peers in the last year?



Did you insult or say nasty or unpleasant things to other children in the last year?



Relations with Friends (Percentage who expressed agreement)

I like to spend time with friends



I have close friends

My friends are always willing to listen

Relations with Parents (in percent)

My parents listen to my problems
I feel closeness to my family
I get along fine with my parents

Peer's Delinquency (in percent)

Break-into and steal from cars
Break-into and steal from apartments
Take things from shops without paying

n = 740

Most of the respondents reported positive family relations. A large proportion of the sample indicated that their parents paid attention to their problems (92%), and that they felt closeness to their families (91%) and got along well with them (87%). Relations of friendship appear to have been stable for a large portion of the sample. Almost 82 percent reported that they liked to spend time with friends, 73 percent had close friends for years, and 59.4 percent thought that their friends were always willing to listen to them.

An important finding refers to the extent of exposure to peers who were involved in delinquency. Almost half of the sample reported that they had peers who had shoplifted, more than a quarter (27.5%) reported that they had peers who had stolen from cars, and 17 percent had peers who had committed residential burglaries. This finding is consistent with Warr’s (1993) argument that by adolescence a high percentage of youth are exposed to delinquent friends. 

The overlap between aggressive behavior and victimization was explored next. As shown in Table 2, while 35.9 percent of the offenders had also been victims of violence, only 4.2 percent of the non-offenders had suffered violent victimization by peers. Thus, a violent respondent was over eight times more likely to be victimized than a non-violent respondent. The association between violent offending and victimization was relatively high as indicated by phi=.423 (p < 0.01). This too casts some light on the finding that the proportion of victimized youth was higher than the proportion of offenders. It shows that victimization was not randomly distributed, and was more prevalent among offenders. Thus, starting a fight might have resulted in the victimization of the offender, and often youth known to be violent became victims even without initiation of violence. On the other hand, respondents with no violent history ran a much lower risk of physical victimization.

Table 2 
Relationship Between
Offending and Victimization  

 Offender Type



Violent offender












Non Violent offender













Verbal offender












Non-verbal offender














*p < 0.01


This trend is also clearly seen in verbal aggression. A large proportion of all the individuals who admitted to verbally victimizing others (58.8%) were themselves victims of such behavior. On the other hand, of all the respondents who were not involved in verbal aggression, only 14.9 percent were victims of such behavior. The association between verbal aggression and victimization is substantial and statistically significant, as indicated by phi=.419 (p < 0.01).

To sum up this point, the descriptive results show an overlap between aggressive offending and victimization. The proportion of offenders who were also victims is significantly higher than the proportion of non-offenders who reported victimization. The overlap between offending and victimization was greater in the case of verbal aggression than in the case of violent conduct. However, in both cases the likelihood of an offender (verbal or violent) becoming a victim was much higher than the likelihood of a non-offender becoming a victim.

Multivariate Analysis

Next, a multivariate analysis was conducted to learn which variables explained verbal and violent victimization. Since the dependent variables in this study were dichotomous, the logistic model was used to estimate the likelihood of the occurrence of verbal and violent victimization (DeMaris 1992). By this procedure one can examine the contribution of each independent variable, to the explanation of the odds of victimization.

The formal logistic model is:

Log Of Odds = Log (p)/(1-p) = a + b1x1 + b2x2 + ….+bkxk + e.

A common interpretation of the coefficients of the independent variables is using the following transformation {100*[odds ratio - 1]} (Demaris 1992; Menard 1995). The result indicates the percentage that each independent variable increases/decreases the odds of the dependent variable.

Table 3 presents the results of a logistic regression predicting the likelihood of an adolescent being a victim of verbal aggression. The literature considers violent conduct to be a critical variable in explaining victimization. Therefore, for simplicity of presentation, to determine the effect of violent conduct, two models are presented. In the first, all relevant independent variables (except violent conduct) were included. In the second, violent conduct was added. By juxtaposing the two models we could evaluate whether the addition of violent conduct improved the fit of the logistic equation to the data.


Table 3 
Results of Logistic Regression
Predicting Verbal Aggression 

Variable Name

















Gender (1=female)














Parent’s relations














Peer’s relations














Peer’s delinquency














School satisfaction














Neighborhood violence














Violent behavior















-2 Log Likelihood







Goodness of Fit












*p < 0.01
**p < 0.05

The results show that the younger adolescents were more likely to suffer from verbal aggression than older adolescents, and males were more likely to be verbally abused than females. Furthermore, the quality of the adolescents’ relations with their parents is related to the likelihood of victimization by verbal aggression. The odds of verbal victimization for adolescents close to their parents are 17.3 percent lower than for adolescents not close to their parents. The findings also suggest that adolescents who reported distant relations with their peers were at a higher risk of being verbally victimized than adolescents who reported good peer relations. The odds of verbal victimization for adolescents close to their peers are fifteen percent lower than for adolescents not close to their friends.

The regression also indicates that involvement in a subculture that tolerates violence was also related to the likelihood of being a victim. In accordance with this finding, adolescents who identified their peers as property offenders and adolescents who reported that in their neighborhood it was common for adults to beat their children were also more likely to be victims of verbal aggression. In addition, and most importantly, as shown in model 2, individuals who were violent toward their peers were also more likely be victims of verbal aggression. The odds of victimization from verbal aggression for adolescents that engage in violent behavior is fifty-eight percent higher than for the ones that do not report violent behavior. The introduction of the variable violent conduct reduced the -2 log-likelihood, indicating that its introduction increased the ability of the model to predict victimization.

A similar analysis was conducted where the dependent variable was violent victimization. The first model shows the results without violent conduct; the second model is identical to the first but with the addition of violent conduct as an independent variable.


Table 4 
Results of Logistic Regression Predicting
Victimization from Violent Behavior 

Variable Name

















Gender (1=female)














Parent’s relations














Peer’s relations














Peer’s delinquency














School satisfaction














Neighborhood violence














Violent behavior















-2 Log Likelihood







Goodness of Fit












*p < 0.01
**p < 0.05

Table 4 shows that age was negatively related to violent victimization. The younger the adolescent, the higher the likelihood of his/her becoming a victim of violent conduct. Females were less likely than males to be victims of violence. In contrast to verbal aggressive behavior, in the model predicting violent victimization the effect of relations with parents and peers was not found to be statistically significant. Consistent with the results presented in Table 3 were the effects of having delinquent friends and neighborhood violence. The odds of violent victimization for adolescents exposed to delinquent friends is 35 percent higher than the odds of adolescents not exposed to delinquent friends. In addition, the odds of adolescents that reside in neighborhoods in which violence toward children is common are 99 percent higher than the odds of adolescents that are not exposed to such an environment. Furthermore, the results show that those who acted violently against their peers during the previous year were also more likely to be victims of violence than individuals who were not involved in any violent conduct against their peers. The introduction of violent offending reduced the -2 log-likelihood, indicating that the inclusion of this variable improved the fit of the model to the data.


The main purpose of this study was to examine the relationship between offending and victimization among an adolescent population. One hypothesis tested here was derived from the literature on bullying. It maintains that the existence of social roles in terms of victims is distinct from perpetrators. Clearly, this hypothesis was not supported by our findings. On the contrary, we found a high association between the aggressive conduct of youth and the probability of their becoming victims of aggressive conduct. In fact, any role differentiation, if one exists, is between those who act violently and also find themselves on the receiving end of such conduct, and those who are neither perpetrators nor victims of such behavior. Apparently, engaging in violent conduct increased the likelihood of also being the recipient of violence. It is overwhelmingly evident that there is a very small likelihood for non-violent actors to be the targets of violence, and therefore it is safe to conclude that non-violent adolescents by no means constitute the pool of victims of violence.

Furthermore, these results offer a very limited support for the power imbalance hypothesis. The expectation that adolescents who are physically and socially powerless are more likely to be victims was only partially supported. Age (the younger are more likely to be victims) seemed to be the only variable associated with physical victimization. However, contrary to the power imbalance hypothesis, females were less likely to be victims than males. With regard to social power, only in the case of verbal aggression did the quality of relations with friends and parents matter. Accordingly, the better the relations of the adolescent with peers and parents, the less likely was he/she to be verbally abused by peers. However, in the case of victimization by physical violence, the quality of the relations with parents and friends was statistically significant. Also, contrary to expectations, adolescents who reported high satisfaction with the school were not prone to being more or less victimized by any type of aggressive behavior, physical or verbal.

The findings support both hypotheses, the subculture of violence hypothesis and the social proximity hypothesis. Adolescents who know delinquent peers and live in neighborhoods where violence against children is common are more likely to be victims of both verbal and physical aggressive behavior. It is important to note that the effect of neighborhood is very strong in both models, increasing the odds of victimization by eighty percent in the case of verbal aggression and by seventy-seven percent in the case of violent victimization. This finding also contradicts any expectations of finding youth specializing in delinquent activity and, hence, involved either in property-related crimes or in violent offenses. It seems that among youth engaged in delinquency, there is no evidence yet of any crystallization of a specialized career pattern. We found a significant correlation between youth belonging to a delinquent subculture, which entailed involvement in theft and property offenses, as well as aggressive behavior, and the likelihood of their physical victimization. A delinquent subculture presumably entails a variety of activities, among them property crime and violence.

The findings suggest that in neighborhoods where violent norms are common, and young males associate with delinquent peers, a subcultural environment seems to exist that supports the use of aggressive behavior. Under these conditions, individuals are very likely to alternate between the roles of offenders and victims (Singer 1981). Perpetrators may become victims because they hold values that support resorting to violence to resolve disputes, and victims may become offenders because of norms that justify and encourage retaliation. 

Top of Page


Baldry, Anna, and David P. Farrington. 1998. "Parenting Influences on Bullying and Victimization." Legal and Criminal Psychology 3:237-254. 

Baldry, Anna and David P. Farrington. 2000. "Bullies and Delinquents: Personal Characteristics and Parental Styles." Journal of Community and Applied Psychology 10:17-31. 

Best, Joel. 1990. Threatened Children. Chicago, IL: University of Chicago Press. 

Best, Joel. 1994. "Troubling Children: Children and Social Problems." Pp. 3-22 in J. Best, editor, Troubling Children, Studies on Children and Social Problems. New York, NY: Aldine de Gruyter. 

Bosworth, Kris; Dorothy L. Espelage and Thomas R. Simon. 1999. "Factors Associated with Bullying Behavior in Middle School Students." Journal of Early Adolescence 19:341-362.  

Boulton, Michael. 1994. "The Relationship Between Playful and Aggressive Fighting in Children, Adolescents and Adults." Pp. 23-41 in J. Archer, editor, Male Violence. London, England: Routledge. 

 Boulton, Michael and Kerry Underwood. 1992. "Bully/Victim Problems among Middle School Children." British Journal of Educational Psychology 62:73-87. 

Cohn, Tessa. 1987. "Sicks and Stones May Break My Bones but Names Will never Hurt Me. Multicultural Teaching 5(3):8-11.  

DeMaris, A. 1992. Logit Modeling: Practical Applications. Newbury Park: Sage Publication. 

Esbensen, Finn-Aage, and David Huizinga. 1991. "Juvenile Victimization and Delinquency." Youth and Society 23(2):202-228. 

Fagan, Jeffrey, Elizabeth S. Piper and Yu-The Cheng. 1987. "Contribution of Victimization to Delinquency in Inner Cities. Journal of Criminal Law and Criminology. 78:586-613. 

Farrington, David P. 1993. "Understanding and Preventing Bullying." Pp.381-458 in Michael Tonry, editor, Crime and Justice, A Review of Research

Finkelhor, David and Nancy L. Asdigian. 1996. "Risk Factors for Youth Victimization: Beyond a Lifestyles/Routine Activities Theory Approach. " Violence and Victims 11(1):3-19. 

Finkelhor, David and J. Dziuba-Leatherman. 1994. "Victimization of Children." The American Psychologist 49(3):173-183. 

Jensen, Gary F. and David Brownfield. 1986. "Gender, Lifestyles and Victimization: Beyond Routine Activity Theory." Violence and Victims 1:85-99. 

Lauritsen, Janet L. and Kenna F. Davis Quinet. 1995. Repeat Victimization Among Adolescents and Young Adults." Journal of Quantitative Criminology 11(2):143-166. 

Lauritsen, Janet L., John Laub, and Robert J. Sampson. 1992. "Conventional and Delinquent Activities: Implications for the Prevention of Violent Victimization among Adolescents." Violence and Victims 7(2):91-108. 

_____. 1991. "The Link Between Offending and Victimization among Adolescents. Criminology 29(2):265-292. 

Meier Robert F. and Terance D. Miethe. 1993. Understanding Theories of Criminal Victimization in Crime and Justice: A Review of Research, 17, pp. 465-495. 

Menard, S.W. 1995. Applied Logistic Regression Analysis. Newbury Park: Sage Publication. 

Olweus, Dan. 1991. "Bully/Victim Problems among Schoolchildren." In . Debra J. Pepler and Kenneth Rubin, editors, The Development and Treatment of Childhood Aggression. NJ :Erlbaum. 

Rigby, K. 1996. Bullying in Schools and What to Do About It. London, England: Jessica Kingsley. 

Roland, Erling. 1989. "Bullying: The Scandinavian Research Tradition." In Delwyin Tattum and David Lane, editors, Bullying at Schools. Stoke-on-Trent: Trentham. 

Salmivalli, 1998. "Intelligent, Attractive, Well Behaving, Unhappy: The Structure of Adolescents’ Self concept and its Relation to Social Behavior." Jounal of Research in Adolescence 8(3):333-354. 

Sampson, Robert J. and Janet L. Lauritsen. 1994. "Violent Victimization and Offending: Individual, Situational and Community Level Risk Factors." In Reiss, Albert J. and Jeffrey Roth, editors, Understanding and Preventing Violence. Washington, D.C.: National Academy Press. 

Singer, Simon. 1981. "Homogeneous Victim-Offender Populations: A Review and Some Research Implications." Journal of Criminal Law and Criminology 72:779-788. 

Sparks, Richard F. 1982. Research on Victims of Crime. Washington, D.C.: Government Printing Office. 

Warr, Mark. 1993. "Age, Peers and Delinquency." Criminology 31:17-41. 

Wells, L. Edward, and Joseph H. Rankin. 1995. "Juvenile Victimization: Convergent Validation of Alternative Measurements." Journal of Research in Crime and Delinquency 32(3):287-307. 

Wolfgang, Marvin and Franco Ferracuti. 1982. The Subculture of Violence. Beverly Hills, CA: Sage Zinnecker, Jurgen. 1998. "Perpetrators of School Violence: A Longitudinal Studyof Bullying in Germany Schools." Pp. 187-203 in Cross-Cultural Perspectives on Youth and Violence. Stamford, CT: JAI Press. 

Top of Page

Gideon Fishman

Gideon Fishman is a professor of sociology and co-director of the Minerva Center for Youth Studies. Currently he is co-principal investigator of a longitudinal study of the social adjustment of adolescent immigrants from the countries of the Former Soviet Union in Israel and Germany. Department of Sociology and the Minerva Center for Youth Studies.

Gustavo Mesch

Gustavo Mesch is a Senior Lecturer in Sociology and Senior Research Associate at the Minerva Center for Youth Studies. He is co-principal investigator of a longitudinal study of the social adjustment of adolescent immigrants from the countries of the Former Soviet Union in Israel and Germany. Department of Sociology and the Minerva Center for Youth Studies. Contact: Department of Sociology, The University of Haifa, Haifa 31905, Israel. E-mail: gustavo@soc. Haifa.ac.il.

Zvi Eisikovits

Zvi Eisikovits is a Professor of Social Welfare and Co-Director of the Minerva Center for Youth Studies. His research interests are domestic violence and Zvi Eisikovits escalation of violence. He is the Principal Investigator of the Israel National Survey of violence against women. School of Social Work and the Minerva Center for Youth Studies. 

Top of Page

Return to Western Criminology Review Archive

Go to Western Society of Criminology Home Page