|Home | About | Journals | Submit | Contact Us | Français|
The No Child Left Behind Act requires state boards of education to identify schools that are unsafe. Schools that are identified by measures such as suspension and expulsion rates are subsequently labeled “persistently dangerous.” To our knowledge there is no published research that attempts to characterize fighting behavior among youth who may attend schools designated as “persistently dangerous.” Two hundred and thirteen sixth grade African American boys and girls attending two urban middle schools on probation for “persistently dangerous” status were examined to investigate differences in demographic characteristics of gender and age and predictor factors of non-parental adult mentorship, parental acceptance of fighting behavior, and peer fighting. These analyses suggest a relationship between the number of peers who fight, youth who believed their parents endorse fighting, and youth without non parental adult mentorship were more likely to fight. This study also indicates that regardless of school status there are modifiable predictors associated with early adolescent fighting.
An important aim of the No Child Left Behind Act of 2001 was to increase order and discipline in schools across the nation. The No Child Left Behind Act (NCLB) requires state boards of education to identify school environments that are unsafe and threaten the health and welfare of children in their school district (U.S. Department of Education, 2003). Among the common indicators used by schools to determine which are unsafe include the number of suspensions and/or expulsions from school for fighting or other interpersonal violence. Middle and high schools that consistently have high levels of violent behaviors for two or three years (depending on the state) are subsequently designated “persistently dangerous,” and suffer penalties, such as voluntary student transfers. Consequently schools at risk for persistently dangerous designation are increasingly interested in strategies that may reduce students' risk for engaging in fighting ad other forms of school violence. To date, there has been limited research on factors which influence fighting behavior in persistently dangerous schools. Having an enhanced understanding of malleable factors which increase or decrease youths' involvement in fights will help target preventive interventions more effectively.
The current study uses data from students at two urban middle schools on probation for persistently dangerous designation to explore three malleable factors which may be targeted through preventive interventions to reduce school violence. Specifically, we explore the influence of adult/youth partnerships, deviant peer behavior, and parent expectations about violent behavior as factors which may influence youths' risk for involvement in fights. In addition to being highly correlated with violent behaviors, these variables were selected because they are potentially malleable factors which can be modified through targeted interventions to meet the needs of youth most at risk for engaging in violence. In the following sections, we consider each of these factors and their relation to youth violence.
Physical fights have been defined as “two or more teens who have chosen to use physical force to resolve a conflict or argument” (National Youth Violence Prevention Resource Center, 2002, p.1). Approximately 33% of high school aged youth fought at least once in the thirty days prior to the study (Grunbaum et al., 2004), whereas studies of middle school students suggest that fighting is even more common (37-60%) (Aspy et al., 2004; Clubb et al., 2001; Cotten et al., 1994; Malek et al., 1998a; Oman et al., 2001). In addition to the immediate health risks of engaging in fights (CDC, 2000), involvement in violence during early adolescence a risk factor for subsequent involvement in other risky and aggressive behaviors (Borowsky & Ireland, 2004; Moffitt, 2006).
Consequently, there has been considerable interest in risk factors for involvement in violence. Quite often, the research has focused on demographic risk factors, such as gender, ethnicity, and age; however, there is increasing interest in predictors which are potentially malleable through preventive interventions. The extant research suggests that non-parental adult/youth partnerships, risky peer behavior, and parental expectations of violent behavior are potential risk factors for engaging in violence at school, but which are potentially malleable through intervention programs, which may in turn reduce the schools likelihood for persistently dangerous designation. In addition, we will examine the non-malleable factors of with-in grade age difference and gender. Below, we consider the research on these five factors.
Having peers who engage in deviant, violent, and aggressive behavior is correlated with an individual engaging in similar behavior (Dishion et al., 1994; Espelage, Holt, & Henkel, 2003; Foney et al., 2002). In addition to the correlation between peer fighting and individual behavior (Espelage et al., 2003), youth perceptions of their friends' behaviors are also highly correlated with and/or predictive of their own behaviors (Griffin et al., 1999; Smith et al., 2001).
Youths' perceptions of their parent's acceptance of fighting to defend themselves or when necessary has been linked with the youths' involvement in fights (Copeland-Linder et al., in press; Orpinas et al., 1999; Malek et al., 1998a). Specifically, Malek and colleagues (1998a) found that fighting was significantly more common among middle schoolers who believed their parents want them to fight if insulted. Similarly, Orpinas and colleagues (1999) found a positive association between students' fighting and students' perceptions of parental support for fighting.
Research on positive youth development has consistently shown that having a positive relationship with a caring adult is a protective factor for involvement in delinquency and youth violence (Aspy, et al., 2004; Rhodes et al., 2002). Some of these relationships are naturally formed, whereas others are fostered through mentorship programs (e.g., Big Brothers/Big Sisters) (Royse, 1998). Youth who reported having an adult whom they admire and with whom they have a relationship, as opposed to youth who reported that their role model was a celebrity or historical figures, have higher grades and self-esteem (Yancey et al., 2002), and were significantly less likely to carry a weapon, use illicit drugs, smoke cigarettes, and have sex with more than one partner (Beire et al., 2000). A recent study by Aspy et al. (2004) found that having a non-parental adult role model was negatively associated with fighting (Aspy et al., 2004). The extant research suggests that positive youth/adult relationships may be an important malleable factor which reduces youth's likelihood for involvement in fights.
Although not widely acknowledged, there is increasing evidence that children who are retained a year or more in school are at increased risk for subsequent aggressive and problem behavior (Nagin et al., 2003). Approximately 9% of children are retained at some point in their educational career (National Center for Educational Statistics, 2006). Most grade retentions in elementary school occur because of educational or social/developmental delays, whereas retentions in middle school occur because of academic problems (National Center for Educational Statistics, 2006). Regardless of the reason, retained students are typically older than their classmates (Malek et al., 1998b) and are likely have existing social or cognitive challenges which increase their risk for involvement in violent and delinquent behaviors (Borowsky, Ireland, & Resnick, 2002; Malek et al., 1998b). Consequently, we consider within-grade age differences to be a risk factor for involvement in fighting.
Gender differences in fighting and other forms of youth violence are commonly reported in the research (Franke et al., 2002; Grunbaum et al., 2004a; Molnar et al., 2005; Yonas et al., 2005). One study estimated that monthly rates of fighting for middle school aged males were approximately 1.6 times the rate for females (35.9% females and 50.2% males) (Clubb et al., 2001). In their study of early adolescent African-American youth, Cotten and colleagues (1994) found that males (47%) were approximately two times as likely to fight as their female (26%) counterparts.
Although there continues to be great interest in risk and protective factors for involvement in youth violence, there has been limited research on malleable factors which reduce the risk for involvement in violence among students attending schools on probation for persistently dangerous designation. Having a better understanding of factors which are potentially malleable may direct our attention to preventive interventions that address these targets. This in turn may ultimately reduce the level of violence within these schools.
The current study examines a set of three potentially malleable factors which involvement in fights. The data come from students at two urban middle schools on probation for persistently dangerous designation. We explored whether non-malleable factors of with-in grade age differences and gender, along with malleable factors of peer fighting, parental expectations of violent behavior, and positive non-parental youth/adult relationships were associated with the youth's fighting.
Data for this study come from the first year of a larger three year randomized controlled trial of a school-based mentoring program for sixth-grade students (N= 477) conducted at two urban middle schools which were on probation for designation as persistently dangerous by the state of Maryland. The only data available for comparison purposes were the race and gender for all students enrolled in participating schools, which is published by the Baltimore City Public School System. When comparing the distribution of the sample of 211 to the total of 1589 students enrolled in these two schools, the sample was not significantly different with respect to race and gender.
Only students who were entering the sixth grade for the first time and were not in self-contained Special Education classrooms were eligible for participation in the larger study. The current sample included 211 sixth graders, from among 448 who were eligible for study enrollment and consent was solicited. A total of 163 (30% of eligible participants) did not return consent forms. Of those who did return the consent forms 44 (9% of eligible participants) refused participation and 241 (54% of eligible participants) agreed to participate. Some of those youth who agreed to participate were excluded from the current analyses because they did not respond to key study variables (n=20) and because of their small sample size students who were not African-American (n=8). Data specific to those who refused to participate are unavailable; however, publicly available school-level data from these two schools suggest that the current sample of 211 is similar to the 1,589 students enrolled in these two schools, with regard to race (95% and 97% African American), gender (47 % and 45% female), and students receiving special education services (17% and 21%). In addition, close to 90 percent of the student body at both schools received free or reduced lunch from the federal lunch program indicating that a high number of students were from low income families.
Recruitment occurred through letters sent to parents and presentations given at school parent orientation nights. Small incentives (e.g. class pizza parties and gift cards) were offered to participating teachers and classrooms for high response rates regardless of the number of consents or refusals. Written parental consent and adolescent assent were obtained for each participant. Only the baseline data are analyzed in the current paper. This study was approved by the Institutional Review Boards at the participating research institutions and the school board.
Youth completed a series of self-report measures in the school computer labs using a Computer Assisted Survey Instrument (CASI). Two 45-minute data collection sessions were conducted to ensure that each child had ample time to complete the measures. The items analyzed in the current study are described below.
A 12-item scale adopted from Orpinas et al. (1999) as utilized elsewhere (Linder et al, in press) were used to measure youth perceptions of parental expectations of fighting behavior under specific circumstances. Students had the choice of agreeing or disagreeing with statements about what their parents or guardian would want them to do in situations such as “Ignore someone if he or she calls me a name,” and “Try to talk my way out of it if someone asks me to fight.” A mean scale score was calculated such that a higher score indicated more aggressive parental attitudes, with a range from 1 to 10 (alpha = .75). Based on the distribution of the responses, the scale was collapsed into quartiles: a low belief (score ≤ 5.42), a middle-low belief (scale score ranging from 5.42 to 6.49), a middle-high belief (scores ranging from 6.50 to 8.08), and a high belief (score ≥8.09).
The following single item was used to evaluate peer fighting behaviors: “How many of your 5 friends get into physical fights with other kids?” (Simons-Morton, 1999). Based on the distribution of the responses, this item was categorized into: no friends who fight (a response of 0), few friends who fight (response of 1-2), some friends who fight (response of 3 or 4), and most friends fight (response of 5).
Having positive relationships with adults, other than ones parents was measured by six items from the National Longitudinal Study of Adolescent Health regarding the influence of non-parental adult relationships. Students responded to six statements about an adult other than their parent or guardian, including there is an adult “With whom I can talk about personal problems,” “Who is concerned with my well being,” or “Who is there for me when I need help.” Participants indicated the extent to which they agreed with each statement on a 10-point scale ranging from 1, strongly disagree, to 10, strongly agree. Mean scale scores were calculated for respondents such that higher scores indicated a more favorable relationship with a non-parental adult (alpha = .88). Based on the distribution of the responses, the scale categorized into higher non-parental youth/adult partnership (scale score of 10) and lower youth/adult partnership (scale score values less than 10). This item was used included in analysis as dichotomous.
The following two items modified from the Orpinas Aggression Scale (Orpinas et al., 1995) were used to measure youth fighting behavior: “In the last 30 days, how many times have you pushed, shoved, slapped, or kicked another person at school?” and “In the last 30 days, how many times have you pushed, shoved, slapped, or kicked another person at home/neighborhood?” Based on the distribution of the responses, these items were combined and dichotomized into those who fight (score values greater than and equal to 1) and those who do not fight at home/neighborhood or school (score values equal to 0).
Data were analyzed in Stata 8.2 for Windows (Stata Corporation, 2002) using univariate analysis and regression procedures. Chi Squares were conducted to determine differences of fighting based on demographic variables and logistic regressions were employed to determine odds of fighting by independent variables. To take into account potential differences by schools, all analyses were controlled for school membership. Similarly, gender was included in all analyses as a covariate. Children who engaged in fights were compared to children who did not fight using simple logistic regression. Next, bivariate relationships between fighting and predictor variables were determined using logistic regression analysis. Third, multiple logistic regression models were used to determine the odds of fighting as related to each predictor variable examined simultaneously (Rosner, 2000).
The outcome of fighting and all factors of the study population are presented in Table 1. Less than half (41%) of the students had engaged in at least one fight at home/neighborhood, school or both places during the thirty days prior to the survey. Of the total population 16% fought at home/neighborhood, 12 % at school, and 14% both places. There were no differences in the odds of fighting between the two schools.
Bivariate evidence supported associations regarding all malleable factors and fighting. These data support that youth who reported lower non-parental adult/youth relationships (a scale score of 10) were more likely to fight than those who had the highest score for non-parental adult/youth relationships (a scale score of 10) (unadjusted OR: 2.22). There was significant evidence that friends who fight had an impact on fighting behavior. Youth who reported having 3-4 fighting friends had increased fighting behavior as compared to those who reported that none of their friends fight (OR: 4.24, p<.05). Furthermore, youth who reported that all five of their closest friends fight, showed increased fighting behavior when compared to those who reported none of their friends fight (OR: 5.11, p<.05). In addition, the importance of youth perceptions of parental expectations of fighting behavior under specific circumstances was associated with fighting behavior. Youth who reported their parents would conditionally accept fighting, fought more than those who reported their parents would not accept fighting in any situation (Low compared to Mid-High OR: 2.66, CI: 1.17-6.03; Low compared to High OR: 2.25, CI: 1.02-4.95).
When adjusted for other youth factors, substance use remained statistically significant. In analysis adjusting for predictor factors only, non-parental adult mentorship, friends fighting, and parent conditional acceptance of fighting remained significant. See Table 2 for adjusted analysis.
A full model was run to include all youth and predictor factors; the results are reported in Table 3. In this full model, many youth and predictor factors were statistically significant covariates, substance use, age, peer fighting, and parental conditional acceptance of fighting. These data did not suggest that there were fighting differences based on gender, but an interaction between non-parental adult mentorship and gender was found. When comparing males to females among youth who reported high non-parental adult mentorship, no fighting differences were found. However, there was a statistically significant difference in fighting between girls and boys reporting lower NPAM. When comparing those who have less adult mentorship, females were more likely than males to fight.
Although research indicates that youth are safer at school than at home or in the community (Malek et al., 1998a), school violence continues to be of great concern for educators, policy makers, and families. The primary objective of this study was to investigate fighting among youth who attend schools on probation for persistently dangerous status. This study provides a snapshot of the youth attending these schools, with the intention of identifying malleable factors which can be influenced through preventive intervention programs.
A large portion of the participants (41%) reported fighting in the 30 days prior to the survey, which indicates that fighting is a common behavior among this population. This is in the range of 30-50% of fighting in the previous 30 days that was reported by other researchers measuring fighting among early adolescents (Clubb et al., 2001; Cotten et al., 1994). We also found that fighting was more common among those who used alcohol, tobacco and cigarettes when compared to those who do not use them. Under aged drinking, smoking and drug use are major risk factors associated with fighting (Lowry et al., 1999, Orpinas, et al 1995). Whether substance use contributes to fighting behavior or that these behaviors co-occur is not clear from this cross sectional analysis. The early adolescent youth in this study live in poor urban neighborhoods where they are exposed to high-violence and illegal drug activity. As these youth get older, their involvement in violence and substance use may increase because of these environmental influences.
Within grade age differences were also examined in this study. Previous research suggested that youth older than others in the same grade were more likely to fight (Malek et al., 1998b; Borowsky et al., 2002). The current study found that students who were 11 years and younger, were more likely to fight than their older classmates. Differences in environment including exposure to violence may explain the differences in these findings and previous research. Research has found that females who live in high violent neighborhoods fight as much as their male counterparts (Howard et al., 2002) in order to establish a reputation that they can defend themselves and to reduce the risk of future fights (Ness, 2004; Jones, 2004). Similarly, younger adolescents who attended these schools may fight more in order to reduce the threat of future fights. Youth who live in high crime, high violence neighborhoods are at greater risk of engaging in violent behavior including fighting (DHHS, 2001; Roche, Webster, Alexander, & Ensminger, 2003). These two schools on probation for persistently dangerous status were located in low-resource, high violence neighborhoods.
Youth were asked about their perceptions of their parents' expectations fighting of behavior. Our results are consistent with these earlier studies (e.g., Malek et al., (1998), Orpinas et al., (1999)) indicating that youth who perceived their parents endorse fighting to resolve conflict were more likely to fight. Ohene and colleagues (2006) examined the relationship between perceived and stated parental expectations regarding adolescent violent behavior. While they found that parental stated expectations did not correlate with youth behavior or youth expectations, they did determine that youth perceptions of expectations were highly correlated with youth violent behavior. Smith and colleagues (2001) found that parents' impact on fighting behavior for African American adolescents was through the pathway of their friends' behavior (Smith et al., 2001). In contrast, the current study suggested that youths perceptions of their parents' attitudes were directly associated with fighting among early adolescent African-American youth fighting behavior. Some researchers suggest that parental warmth (Stattin & Kerr, 2000; Fletcher et al., 2004) and parental knowledge of youth behavior (Stattin et al., 2000; Pettit, Bates, Dodge, & Meece, 1999; Steinberg, Fletcher, & Darling, 1994) are most important. Others suggest that parental school involvement (Dornbusch, Ritter, Leiderman, Roberts, & Fraleigh, 1987) and parental monitoring (Steinberg, Lamborn, Darling, Mounts, & Dornbusch, 1994) are key in reducing the number of youth engaging in risky behavior.
Although researchers have yet to explain how parental involvement influences youth behavior, there is overwhelming support for the idea that youth perceptions of their parents are key in improving youth adjustment and prosocial behaviors while reducing a child's participation in risky behaviors including fighting (Dishion et al., 2002; Fletcher et al., 2004; Orpinas et al., 2000; Steinberg et al., 1994).
Another interesting finding in this study is the impact of peer behavior on youth fighting behavior. The current data support existing research that having peers who engage in violent and aggressive behavior is correlated with an individual engaging in similar behavior (Espelage et al., 2003; Foney et al., 2002; Griffin et al., 1999; Smith et al., 2001). In our study, fighting increased as the number of peers who fight increase. Peers should be considered in any intervention aimed at reducing fighting behavior.
Lastly, these data were analyzed to determine the associations between predictor factors and fighting. In this analysis all predictor factors were found to be associated with fighting. Early research has determined that mentoring is beneficial for youth development, especially those from low income families and single parent households (Yancey et al., 2002; Royse, 1998; Rhodes, Bogat, Roffman, Edelman, & Galasso, 2002). Many youth in this study identified themselves as having adult mentorship and these youth engaged in less fighting behavior than youth who did not report high mentorship. Girls who live in low income high violence neighborhoods may benefit most from adult mentorship.
Methodological limitations should be noted when interpreting study results. This study relies solely on a single instrument measuring youth self-report, which has the potential for social desirability bias creating systematic error. Although there are threats to validity based on the methods used to measure youth aggression, this study examined self-reported violent behaviors because these behaviors have not decreased in number as have other violent acts such as weapon carrying, homicides, and youth arrests (DHHS, 2001). Additionally, rigorous pilot testing of the measures and confidentiality of survey administration allow for greater confidence in the validity of the data. These data collection methods are also consistent with state-of-the research methods in youth violence research.
There are limitations in this study based on the use of a cross sectional design. Cross sectional research can suggest associations between two constructs but cannot determine causality. The small sample for this study was not randomly chosen; all 6th grade students in the selected schools were eligible to participate. The schools in this study were chosen based on convenience and need for the intervention. Results may not be generalizable to youth in other settings.
Despite these limitations, the current study adds to the small body of literature on risk factors for fighting among youth attending persistently dangerous schools. This research suggests there are environmental influences that should be considered when examining fighting and other violent behaviors.
This study indicated that there are a significant number of youth do not fight, even when they attend schools suggested to have an increased risk of violence. The regression analyses in the current study were common malleable factors indentified among youth who refrained from fighting. For example, future interventions focused on altering parental attitudes towards youth violence and adolescent's perceptions of their parents attitudes towards violence could have the greatest impact on youth fighting, and may reduce the levels of violence in schools.
In addition to parental support and communication, mentoring programs are often used to enrich child and adolescent development. Currently, programs like Big Brothers, Big Sisters are attempting to provide mentors to youth, mostly from low income families. Unfortunately, these programs often have waiting lists, are unable to match children, or make matches that are not successful. An unsuccessful adult/mentor match could result in negative youth outcomes because the youth could feel let down or abandoned by their mentor (Rhodes et al., 2002). It is important to point out that many youth successfully identify mentors without the assistance of these programs. They have adults in their lives that they see as caring and supportive. These youth have skills to independently identify mentors and they experience the same positive effects as youth in mentoring programs. One approach that researchers and program developers could take is to help more youth identify and bond with those mentors already in their lives.
Vanya C. Jones, Johns Hopkins University Bloomberg School of Public Health, 200 N. Wolfe Street, Suite 2093, Baltimore, Maryland 21287.
Cathrine P. Bradshaw, Johns Hopkins University Bloomberg School of Public Health, 624 N. Broadway, Hampton House, suite 831, Baltimore, Maryland, 21205.
Denise L. Haynie, Prevention Research Branch, DESPR, NICHD, NIH, 6100 Executive Blvd. Rm. 7B13 MSC7510, Bethesda, Maryland, 20892.
Bruce G. Simons-Morton, Prevention Research Branch, DESPR, NICHD, NIH, 6100 Executive Blvd. Rm. 7B13, Bethesda, Maryland, 20892.
Andrea C. Gielen, Johns Hopkins University Bloomberg School of Public Health, 624 N. Broadway, Hampton House, suite 554, Baltimore, Maryland, 21205.
Tina L. Cheng, Johns Hopkins University School of Medicine, 200 N. Wolfe Street, Rm 2055, Baltimore, MD 21287.