Teen Violence: Examining the Effect of
Changing Schools Frequently

Victoria Peters

Thousands of children every year must enter and exit elementary, junior, and high schools. They must leave friends, find new ones, enter into extracurricular activities, and simply try to fit in. Many times we hear horrific stories of gun and weapon violence and fighting between students in these situations. Why is it that the ďnew kidĒ is always the one who is violent? They may be unattached from school activities or have no commitment to any goals. There may be an underlying issue from becoming predisposed to violence at home. Although mobility and social capital has been investigated (Homans, 1961), violence causes violence regardless of the situation. The compelling questions of why those children who are newer to the school are more prone to violenct acts have not been answered.

Findings from a cohort study show that being abused or neglected as a child increases the risk of violent criminal behavior, but have those children been attached to the community long? School violence has become a serious problem in recent decades, especially where weapons such as guns or knives are involved (Widom, 1989). This also includes violence between school students, as well as physical attacks. A basis can be built from social disorganization theory which refers to the failure of the social institution in communities and neighborhoods. However, this theory has not investigated escalated violence outside of those neighborhoods. 

Hirschi's Causes of Delinquency (1969) is a benchmark for theory construction and research in the delinquency field. By using Hirschiís social bonding theory, specifically focusing on commitment and attachment, I will try to uncover why mobile children start fights, bring weapons to school, or seem to be isolated or engage in violent behaviors. I hope to discover if this is because they are experiencing an unstable home environment where they have to pick up and leave to start a new life in a foreign area. I will attempt to broaden the horizons of Hirschiís four elements of the bond by providing answers to why school related violence exists in communities every year.



            As noted above, Hirschiís Bonding Theory (1969) has been involved in many areas of social science study. Most commonly, scientists have used his four elements of the bond; attachment, involvement, commitment, and belief. In this theory, Hirchi was striving to provide an explanation for juvenile delinquency. As he sees it, ďÖthe delinquent person is relatively free of intimate attachments, the aspirations, and the moral beliefs that bind more people to a life within the law.Ē (Hirschi, 1969, preface). In his social control theory of delinquency, Hirschi identifies four elements of the bond, and assumes that if the individualís bond is broken to society they will result in a delinquent act. Attachment is explained through means of a psychopath; in other words, the deficient attachment to or affection of others results in violent acts. Less attached individuals are free from moral constraints and guilty consciousness. Commitment is defined through the concept that most people live in organized society, acquire goods, reputations, and prospects that they do not want to risk losing (Hirschi, 1969). Thus, the person is committed to a conventional means of living. Involvement includes the level of engrossment an individual has towards a conventional activity. The assumption is that a person may be too busy doing conventional things to find time to engage in deviance. Finally, belief assumes that there is an existence of a common value system with the society. It is important to note that with belief, we not only assume the deviant has believed the rules, we assume he believes the rules even as he violates them (Hirschi, 1969).

            In general, the more closely related a person is to conventional society, the more closely he will be related in other elements. In research, Hirschiís theory can be combined into six possible combinations. For purposes of the idea of escalated violent behavior as a result of frequently changing of schools, the relationship between attachment and commitment is most prevalent. According to delinquency research, one of the adolescentís problems is the inability to sever ties with parents and peers, thus preventing him from engaging in school related activities and academics (Hirschi, 1969). In relation to Hirschi, his attachments are getting in the way of his conventional commitments. In stratification research, the males who break away from these attachments are more likely to become more mobile thus change homes and schools more frequently.


According to a model used by Nancy Cunningham, behavior will be pro-social or antisocial depending on the ďpredominant behaviors, norms, and values held by those individuals or institutions to which/whom the individual is bondedĒ (Catalano et al., 2004, p. 252) and emphasizes the development of pro-social bonding as a protective factor against harm. Once the bonds are strongly established, they inhibit behavior consistent with the norms and values of the institution to which they are bonded (Cunningham, 2007). In this research, Cunningham used two scales of bonding for attachment and commitment; while attachment consisted of six scale items and commitment consisted of eight scale items. High scores on the scales were identified as a higher attachment and commitment to the school. This study relates to mine because I am also focusing on Hirschiís two elements of attachment and commitment in relation to school violence.

            Behaviors can also vary across cultures. For example, if I was investigating with a variable for multiple cultures I would be able to apply Hirschiís theory of social bonding. In a 2007 study by Miyuki Fukushima, Susan Sharp, and Emiko Kobayashi, there was an investigation of social bonding across two cultures; American and Japanese. An argument was developed that claimed that collectivism in Japanese society generates stronger social bonds, than the more individualistic American society, thus explaining a lower level of deviance in Japan. This study was conducted using college students on measures of deviance and social bonds to the school and fellow students. In the end, the results indicated that Japanese students engaged in significantly less deviance than Americans (Miyuki Fukushima, 2009). As stated above, this does not specifically link with my basis, although it is a more current evaluation of Hirschiís theory in more than one culture.

            Hirschiís social control theory was previously used in a similar way I am choosing to do. In a study conducted by Michael Wiatrowski, David Griswold, and Mary Roberts, they formed a basis from Hirschiís social control theory by proposing that delinquents fail to form or maintain a bond to society consisting of attachment, commitment, involvement, and belief (Wiatrowski, 1981). By using data from the Youth in Transition Study, the report developed multivariate models of social control theory simultaneously with how the four elements of the bond operate in relation to delinquency. Delinquency was measured using an index scale of up to twenty-six items adapted from Goldís (1966) self-report measure of delinquency. Among the areas of measurement were; theft, vandalism, interpersonal aggression, delinquency in school, frequency and seriousness of delinquent behavior, and trouble with parents. These variables are similar to the index scale of violence I will be using in my own research study.

It was noted in the aforementioned study that Hirschi does not consider how his four elements of the bond might act simultaneously in affecting the likelihood of delinquent behavior. In addition, instead of analyzing relationships among elements, he instead hypothesizes the relationships between attachment and commitment, commitment and involvement, and attachment and belief. As a consequence, this theory creates questions to be answered. First, the extent to which Hirschiís four elements represent specific components of socialization is not clear. Second, education and occupational aspirations are discussed within Hirschiís theory, but he fails to construct figures such as socioeconomic status. Research has proven that the aspirations of education and occupational status are important in the development of a complete theory towards attachment and commitment (Sewell, 1969; Haller, 1973).


            The relationship between crime and communities has always been a focal point for sociological investigation. In my research, I am also investigating the reciprocal relationship between violent crime and residential stability in neighborhoods. The majority of sociological investigations limit themselves in the degree of this reciprocal relationship. It has been shown that we need to understand not only how violent crime impacts neighborhood residential stability, but how changes in the stability affect crime simultaneously. In a 2010 study, Lyndsay Boggess and John Hipp attempted to test whether the form of stability matters by comparing two types of stability; traditional and novel. They also examined whether the racial/ethnic composition of the neighborhood mattered to the violent crime relationship. After using data from Los Angeles from 1992 to 1997, their results indicated that violent acts of crime increases the path towards residential stability (Hipp, 2010). This type of study would be helpful in my demographic variable testing, thus it could be determined if mobility affects crime with my age group of adolescents.

            Two researchers in 2005 were driven by a concern with adolescent violence and the growing research on the social factors linked to youth violence. One risk factor they chose to investigate was residential mobility. Dana Haynie and Scott South (2005) identified four broad categories of explanatory factors that have previously invoked theorists on residential mobility and adolescent problem behavior: parent-child relationships, psychological distress, experiences of victimization, and peer social networks. Parent-child relationships are a reason why residential mobility might increase violence because moving negatively affects parentsí relationships with their children, thus the children are inadequately supervised and less monitored. As noted, a change of dwelling, community, or school may cause many connections to be severed in the childís network, thus resulting in a loss of social capital (Coleman, 1988). In turn, the less social capital along with residential mobility the adolescentís development is negatively impacted, which can increase a risk in engaging in violent behavior. An important variable in finding social capital is the adolescents stable network structure (Hirschi, 1969). As proven by Travis Hirschi, when adolescents are firmly connected to family, school, etc. it is more difficult for them to misbehave.

A second factor is psychological distress. In this approach, it is considered that the adolescent has been through a substantial amount of confusion in establishing self-image, finding supporting friends and peers, and beginning their emotional and psychological formation (South, 2005). Residential mobility has been characterized as a stressful life event for adolescents. Psychological distress might lead adolescents to develop a ďnothing to loseĒ attitude towards violent activities (Harris, 2002). Victimization may also be a possible explanation because newcomers sometimes encounter personal attacks and are more vulnerable. Higher rates of victimization might lead residentially mobile adolescents to subsequently engage in violent behavior for purposes of defense, retaliation, or to eradicate strains induced by victimization (South, 2005). The final factor in explaining residential mobility on adolescent violence is the difference between peer social networks of mobile and non-mobile youth. Mobile youth report having fewer close friends and less personal intimacy with the friends they do have (Vernberg, 1990) and they are less likely to be the center of peer social networks (South S. J., 2004). In the end, their findings implied that relatively deviant adolescent school networks may be particularly welcoming to new members and that once mobile adolescents become embedded in such network they will tend to adopt the violent behaviors of their members (South D. L., 2005). This finding is the same that I seek to find in my study of school mobility and escalated violence in adolescents.

Combined Theory and Behavior

            The Council on Alcohol and Drug Abuse took Hirschiís social control theory and social violence and bundled them into one research study; similar to my own. It is important to note that this study evaluates the Columbine High School massacre of 1999. As noted previously, Hirschi (1969) identified four elements of the social bond to explain why some juveniles resort to delinquency. Attachment refers to an individualís sensitivity and empathy for the feelings of others. School shooters are angry with those who bullied, belittled, or ostracized them for a period of time leading up to the attack (Vossekuil, 2006). The second element of commitment pertains to the time and energy one spends pursuing a specific activity as getting an education or career (Hirschi, 1969). School shooters do not anticipate getting away with the crime because the shooter has adopted a kill or be killed mindset (Vossekuil, 2006). The third element, involvement, operates with a premise that a person is engrossed in conventional social activities that the individual will not have the time or desire to participate in delinquent acts (Hirschi, 1969). According to Vossekuil, Reddy & Fein (2000), more than 50% of the perpetrators developed a plan of attack at least two weeks prior to the attack. The final element of belief is defined as an individualís belief in social rules and laws of society (Hirschi, 1969). Revenge was a large motive of the school attacks because the perpetrators believed that retaliation was the only option available to compensate the perceived actions of other students (Vossekuil, 2006).

            Research pertaining to peer harassment, bullying, and victimization has captured attention of researchers who understand symptoms of and risks associated with physical and verbal peer aggression. In addition, theories suggesting residential mobility as an important factor in the aggression have been investigated. In this study, I plan to extend the existing literature on social control and mobility in relation to escalated violence in adolescents. I will be including variables such as parental past issues with the law and an eight item violence scale. As discussed in previous literature, I will provide more current information on the issue of escalated teen violence and school mobility.


            The data for this study came from a sample of 13,349 students in grades 6th through 12th in a large public school in the most populous county in Kentucky (Clayton & Wilcox, 2001). In the spring of 1996, a survey was designed to question students about a broad array of smoking, drinking, drug-taking, and other delinquent behaviors. In addition, it measured family background, attitudes toward school, involvement in school victimization and violence, peer behavior and religious attitude.  It was administered by teachers across 22 schools in the form of a written questionnaire and given during one class period for about 45 minutes. Passive parental consent and Institutional Review Board (IRB) approval had been obtained; studentís consent was obtained at the time of administration. Of the 13,255 completed surveys from the students of 21 schools, 1,000 cases were deleted due to missing data on nominal-level measures. In the end, the total useable sample was 12,343. Yet, because of space constraints presented by the student version of SPSS a final sample of 1,500 cases was randomly selected. The sample is relatively unbiased as to socioeconomic demographics. As one would expect, given the small number of cases lost to missing data, the final subsample of cases analyzed is similar to the total number of students in grades 6th through 12th surveyed (Clayton & Wilcox, 2011).

Independent Variables

Changing Schools

             To assess the frequency of changing schools, the question asked the number of times the respondent had changed schools in the last twelve months. The response was open-ended for respondents to fill in the actual number.

Dependent Variables

Level of Violence

            To measure overall level of violence, I developed a scale to measure the depth of the violent acts. The violence scale consisted of eight items created into an additive scale and contained a wide range of behaviors from shoving or tripping someone to using a weapon in a fight. Moreover, respondents were asked to answer a yes (1) or no (0) to each question. The scaleís range runs from 0 to 8. Zero indicates no violence and 8 represents the most violence


            In order to ensure that any relationship between mobility and the outcome of escalation in violence were not spurious, I included control variables for the analysis. Specifically, I controlled for demographic differences across students that might explain differences, including age, sex, and race. Sex and race are measured by dichotomous variables (male = 1, female = 0, white = 1, non-white = 0). Age is measured by a continuous variable of actual age in years.




This study attempted to examine the relationship between the number of times an adolescent has changed schools in the past twelve months and their level of violence. I sought to answer questions on why mobile children start fights, bring weapons to school, or seem to be isolated or engaged in violent behaviors.

Bivariate Analysis

Table 2 illustrates the bivariate relations among the study variables. Of particular interest is the significance between the number of times changed schools in the past twelve months and the violence scale. Moreover, significant relationships were shown between the number of times changed schools and race, as well as age. Both of these correlations were negatively significant. In fact, the only variable to not show a significant relationship with the mobility variable was the sex of the individual.

Multivariate Analysis

To test the effect of the violence scale on the number of times changed schools in the past twelve months, while controlling for sex, race, and age, I used an ordinary least squares (OLS) regression model for my outcome variable. As seen in Table 3, the number of times changed schools in the past twelve months was significantly related to the individualís level of violence. For every unit increase in the changing of schools, there is a corresponding increase of .190* in the adolescentís level of violence, regardless of sex, race, or age. Compared to females, males score on average 1.555* higher on the violence scale, regardless of race or age. Compared to non-whites, whites score on average .819* lower on the violence scale, regardless of sex or age. For every year increase in age, there is a corresponding increase in a childís level of violence by .072, regardless of sex or race.


            As hypothesized, the number of times an adolescent changes homes does in fact have an effect on their level of violence (See Table 2 and 3). As you many recall the previous information and data does in fact support this hypothesis. It is important to note that this study included many strengths, weaknesses, and limitations.


            In the methods section, I described the creation of a multivariate scale to measure the depth of violence acts among the adolescents. The violence scale consisted of eight items containing behaviors from shoving or tripping to using a weapon in a fight. By doing so, I created a wider range of a single variable to investigate against my independent variables, as well as making their level of violence a more valid measure.


            Due to the fact that this project was based on secondary analysis, I did not have access to all of the variables that ideally should have been included. In addition, limitations of my statistical program provided me with a random sample of 1,500 cases of data; rather than 26,000. Most importantly, this data is only representative of students in this specific Kentucky school district and not of the entire state, nor the nation as a whole. To note, the Kentucky school surveyed was a fairly homogenous state of 87% white citizens.

Directions for Further Research

            As a result of the small sample size, further research should seek to replicate my findings with a much larger sample size. Despite such discussed limitations, I have provided evidence that the number of times an adolescent changes homes in the past twelve months is in fact related to a wide range of violent behaviors. This offers support for Hirschiís social bonding theory (1969).





Catalano, R. F. (2004). The Importance of Bonding to School for Healthy Development: Findings from the Social Development Research Group. Journal of School Health, 252-261.

Coleman, J. S. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, S95-S120.

Cunningham, N. (2007). Level of Bonding to School and Perception of the School Environment by Bullies, Victims, and Bully Victims. The Journal of Early Adolescence, 457-478.

Gold, M. (1966). Undetected Delinquent Behavior. Journal of Research in Crime and Delinquency, 27-46.

Haller, W. a. (1973). Status Attainment Processes. Sociology of Education, 51-91.

Harris, K. M. (2002). Evaluating the Role of 'Nothing to Lose' Attitudes on Risky Behavior in Adolescence . Social Forces, 1005-40.

Hipp, L. N. (2010). Violent Crime, Residential Instability and Mobility: Does the Relationship Differ in Minority Neighborhoods? J Quant Criminol, 351-370.

Hirschi, T. (1969). Causes of Delinquency. Berkeley and Los Angeles: University of California Press.

Homans, G. C. (1961). Social Behavior as Exchange. London.

Michael D. Wiatrowski, D. B. (1981). Social Control Theory and Delinquency. American Sociological Review, 525-541.

Miyuki Fukushima, S. F. (2009). Bond to society, collectivism, and conformity: A comparative study of Japanese and American college students. Deviant Behavior, 434-466.

Pamela Wilcox, R. C. (2001). A Multilevel Analysis of School-based Weapon Possession. Justice Quarterly, 520-522.

Sewell, W. H. (1969). The Education and Early Occupational Attainment Process. Americal Sociological Review, 82-92.

South, D. L. (2005). Residential Mobility and Adolescent Violence. Social Forces, 364-375.

South, S. J. (2004). Friendship Networks of Mobile Adolescents. Social Forces, 315-350.

Vernberg, E. M. (1990). Experiences with Peers Following Relocation During Early Adolescence. American Journal of Orthopsychiatry, 466-72.

Vossekuil, B. R. (2006, October 26). An Interim Report on the Prevention of Targeted Violence in Schools. Retrieved from United States Secret Service: http://www.treas.gov/usss/ntac/ssi_interim_report.pdf

Widom, C. S. (1989). The Cycle of Violence. Science 244.



Table 1. Descriptive Statistics for Study Variables




Descriptive Statistics






Number of Times Changed
Schools in Past 12 Months


(number of times)




Violence Scale

(0=no, 8=yes)







(0=female, 1=male)








(0=non-white, 1=white)





(number of years)












Table 2. Zero-Order Correlations among Study Variables






1. Number of Times Changed Schools in Past 12 Months





2. Violence Scale





3. Sex





4. Race





5. Age






* p < .05


Table 3. OLS Regression Models of Study Variables



Violence Scale

Coeff.           S.E.

Independent Variables


Number of Times Changed
Schools in Past 12 Months




*p < .05

 .190*            .061


1.555*            .131

-.819*            .185

 .072*            .033