Family, School, Community, and Economic Factors Associated with Juvenile Crime in North Carolina: A System Impact Assessment

Nature and Extent of the Juvenile Crime Problem

Rarely does an evening pass in which the locally televised nightly news does not provide coverage of at least one shocking and disturbing act of criminal violence involving juveniles. Nightly reports of drive-by shootings, teens as young as 14 being remanded to adult court to stand trial for murder, gang initiation ceremonies involving sadistic and brutal acts of rape, and caches of drugs and weapons being confiscated from elementary school lockers have permeated the mass media at an alarming and accelerated rate. Unfortunately, these media portrayals are not always anomalies or isolated incidents of shock journalism. Disturbingly these reports can be heard on any big city newscast or read in any small town newspaper. Perhaps even more disturbing is the fact that an even greater number of less sensationalistic, yet nonetheless equally violent, juvenile crimes never make the headlines or front page at all.

A large proportion of this juvenile violence is being perpetrated upon the offenders' own classmates, neighbors, and members of their own age cohort. Snyder and Sickmund (1995) report that results from the 1991 National Crime Victimization Survey reveal that nearly half of the juvenile violent crime victims were attacked by another juvenile. Violence has infiltrated the Nation's youth subculture with teens often preying upon each other as a means of simply understanding, surviving, and coping with the stress, tension, and anxiety that often accompany the troubled period of adolescence.

As Elliott (1993) notes, the most pronounced difference between the youth crime of today, and juvenile crime of the prior decade, is its lethality. Adolescent homicide rates have doubled since 1988 and in 1991 homicide became the second leading cause of fatal injuries among youth under the age of 20 (Elliott, 1993; Snyder and Sickmund, 1995).

The increase in youth violence has dramatically affected and altered the nature and procedural operations of the juvenile justice system; a system which was previously overburdened and often stretched to capacity by its efforts to effectively manage the historically predominant, and less benign, juvenile property offender. While property related offenses continue to constitute the largest percentage of the system's work activities and caseload patterns the rise in juvenile violent crime has exerted a considerable impact upon, and consequently increased, the volume and complexity of law enforcement, juvenile court, and juvenile detention agencies' workloads. Caseload and operational statistics quantitatively demonstrate the impact of increasing violent crime on each of the three major components of the juvenile justice system.

Nationally; the violent crime arrest rate, for youthful offenders under the age of 18, increased 44.8% from 1988 to 1992. A slight decline of 0.65% in the rate of juveniles arrested for property offenses contributed to a modest 1.6% increase in the overall juvenile arrest rate between 1988 and 1992 (Flanagan and Maguire, 1990; Maguire and Pastore, 1994). The rate of petitioned delinquency cases, filed in the Nation's juvenile courts, grew by 24.8% from 1988 to 1992. The largest proportion of this total growth can be directly attributed to a significant rise in the rate of petitioned violent crime cases. The filing rate for petitioned violent crime delinquency cases experienced a 62.7% increase; an increase nearly three and one-half times larger than the 18.2% increase which occurred in the filing rate, for property offense petitions, during this same period (Butts, Snyder, Finnegan, Aughenbaugh, Tierney, Sullivan, and Poole, 1995).

Increases in the arrest and petition filing rates have exercised an influence on the Nation's public juvenile confinement facilities. The juvenile confinement rate expanded by 4.4% from 1987 to 1991. The composition of this confined population also experienced a transformation during this period. The percentage of the population being detained for offenses against persons grew from 25% to 31%; while the percentage of the population being detained for property offenses declined from 44% to 36% (Flanagan and Maguire, 1992; Maguire and Pastore, 1994).

The escalating national trends, in the level of youth violence, and their detrimental consequences, have adversely affected North Carolina's Juvenile Justice System as well. The state's law enforcement, juvenile court, and juvenile detention organizations have not been immune to this increase with each encountering significant growth and change in the quantity and quality of their respective workloads.

The state's juvenile arrest rate grew by 27.3% between 1988 and 1992 and has grown 54.4% from 1988 to 1994. The arrest rate for violent crime climbed 73.3% from 1988 to 1992 and 97.5% from 1988 to 1994. The arrest rate for property offenses experienced a slight 2.3% increase from 1988 to 1992 and an 8.2% increase from 1988 to 1994 (North Carolina Department of Justice, 1989, 1993, 1995). Delinquency petition filing rates increased 13.7% from 1988 to 1992 and 34% from 1988 to 1994. The number of children making an initial appearance before the juvenile court increased 28.3% from 1988 to 1994. The number of juvenile intake complaints in 1992 exceeded the number filed in 1988 by 12.7% (North Carolina Administrative Office of the Courts, 1988, 1992, 1994). Data for fiscal year 1993-1994 reveal that 27% of the teens who were admitted to the state's juvenile detention centers were property offenders and an almost equal percentage (25%) were admitted for crimes against the person (North Carolina Department of Human Resources, 1994). The admission rate for the state's training schools has undergone a 6% increase since 1992 (North Carolina Office of State Planning, 1993).

Causes and Correlates of Juvenile Delinquency

Extensive etiological research on juvenile crime and delinquency has been conducted for decades by an innumerable number of scholars from such diverse disciplines as sociology, psychology, economics, public health, social work, and urban planning. Indeed the modern academic discipline of criminology derives its theoretical roots from these eclectic studies on the causes and correlates of juvenile delinquency. Despite the plethora of prior research a lack of consensual agreement still persists surrounding the exact causes of delinquency, how these causative factors interact with each other, and to what extent the proposed causes can be alleviated or even mitigated by prevention and intervention efforts. A greater degree of consensus can be reached when discussing these same factors within the context of a correlational, as opposed to a causative, role and how to best direct prevention and intervention strategies on the social factors which are associated with the presence of juvenile crime. The following sections will briefly delineate some of the most widely recognized, heavily researched, and extensively debated factors which have been postulated to be either causes or correlates of juvenile delinquency.

Familial Factors

Research on the role and influence of the family, in shaping a child's subsequent behavior, has generally been directed at examining the parenting process and\or the structural and organizational composition of the home. Loeber and Stouthamer-Loeber (1986) reviewed approximately 300 family and delinquency studies and concluded that the greatest predictors of future delinquency were parental supervision, parental rejection, and parent-child involvement. Marital relations, parental criminality, parental discipline, and parental absence were also identified as having moderate levels of influence on a child's subsequent behavior.

Research reports on the effects of parental abuse, and the specific consequences that abusive disciplining techniques exert upon the level of a child's antisocial and aggressive behavior, have produced mixed results. Wright and Wright (1994) succinctly summarize this issue by asserting that abuse is linked to an increased risk of engaging in delinquent behavior; however the underlying mechanisms of this risk have yet to be fully established with a need for more research, on the frequency, duration, intensity, and termination of abuse, being apparent.

Research findings on the relationship between broken homes and delinquency have also generally been inconclusive with some studies suggesting a strong positive relationship between single parenthood and escalating delinquency and others suggesting that the effects of broken homes are complex, non direct, and mediated by numerous other factors such as poverty and community disorganization (Loeber, Stouthamer-Loeber, Van-Kammen, and Farrington, 1991). The link between marital discord and delinquency has been consistently demonstrated with the frequency, duration, and intensity of this disharmony moderating the extent to which this behavior functions as an adequate predictor of juvenile delinquency (See Wright and Wright, 1994, for an excellent summary of the literature surrounding family life and delinquency).

Demographics and Community Structure

Kornhauser (1978) comments that the demographical attributes of age, gender, and race are the most frequently cited and heavily researched correlates of crime. The explanatory power of age and gender persists across time and location with the level of reported crime increasing in direct proportion to the number of young males in the population (Wilson and Herrnstein, 1985). Studies examining the relationship between race and crime repeatedly report an over representation of minorities, especially African-Americans, within the adult criminal justice system. African-Americans are over represented by a factor of four to one among persons arrested for violent crime and three to one for property crime arrests (Wilson and Herrnstein, 1985). This over representation is also apparent throughout each stage of the juvenile justice system with the most pronounced racial disparity being found at the initial intake and detention stages of case processing (Pope and Feyerherm, 1991). Minority over representation in public juvenile confinement facilities remains problematic with the custody rate for African-Americans being nearly five times larger than the custody rate for Whites (Snyder and Sickmund, 1995).

Community structure and the degree of organization within the community have been postulated to be significant factors when explaining the variation between differing communities and their respective crime rates. Increasing urbanization, high population density, excessive residential mobility, and low levels of attachment to the neighborhood have been offered as attributes which may exacerbate the level of crime within a community (Hawkins, Catalano, and Miller, 1992). Poor housing, overcrowded living quarters, physical deterioration, and an absence of natural surveillance by the residents have also been shown to be associated with childhood conduct disorders and delinquency (Bursik and Webb, 1982; Murray, 1983).

Other factors, such as poverty, may interact with these negative community features and consequently produce numerous public health problems for the community and its residents, especially its children. Without financial resources babies may be exposed to more pre and perinatal complications, thus infant and child death rates may rise under these adverse conditions. The absence of strong community bonds, in conjunction with isolation from the legitimate labor market, weakens the community's ability to resist the infiltration of illegitimate drug distribution networks and gangs. Community residents may encounter greater opportunities to participate in this illicit economy, and\or affiliate with these gangs, thus potentiating a possible increase in the amount of systemic violence and murder within the community (Elliott, 1993; Blumstein, 1995).

Academic Performance and the School Environment

Gottfredson (1981) reports an inverse relationship between intellectual ability and delinquency with the failure to meet academic expectations generating conflict, frustration, and stress within the affected students. Consequently the likelihood of engaging in violent delinquent behavior is intensified ( Elliott, 1993). This likelihood is even more pronounced among those students who have been exposed to the use of violence, in the home and community, as an appropriate means of conflict resolution.

As Elliott (1993) asserts academic underachievers and troublemakers are often unintentionally grouped together in the same classroom as a result of the school's emphasis on nurturing and clustering the academically superior students into the same educational track. Delinquent gangs and peer groups often emerge under these circumstances when individual feelings of alienation, frustration, and anger are mutually reinforced within the classrooms of the academically poor students. This collective adaptation to school failure and the mutual sharing of negative emotions often manifests itself in the form of violent group behavior within, as well as outside of, the school environment.

Economic Indicators

Initial findings, from the Office of Juvenile Justice and Delinquency Prevention's longitudinal study on the causes and correlates of juvenile delinquency, reaffirm the link between delinquency and poverty. Youth who reside in underclass areas face a significantly greater risk of engaging in delinquent behavior irrespective of race and the quality of parental supervision (Huizinga, Loeber, and Thornberry, 1994). As previously noted extreme economic deprivation can also exert an additive effect on many of the previously discussed factors such as neighborhood disorganization and academic failure. Thus; these problems become exacerbated when occurring in areas characterized by extreme poverty and high unemployment rates (Farrington, 1991). Poverty appears to function as a critical nexus, or catalyst, for a multitude of other behavioral problems, problems which are also significantly correlated with crime and delinquency. Children who display antisocial and problematic behavior, as well as reside in economically deprived communities, are more likely to develop problems with school dropout, teen pregnancy, alcohol and substance abuse, and future adult criminality (Hawkins, Catalano, and Miller, 1992).

This paper presents the findings of an exploratory study, conducted by the North Carolina Criminal Justice Analysis Center, which sought to identify the family, school, community, and economic factors that exert a significant impact upon the state's juvenile justice system's workload and case processing activities. Specifically, the study sought to isolate the most salient or significant factors associated with juvenile arrests, court filings for delinquency petitions, and training school admissions. The identification of important family, school, community, and economic indicators will enable law enforcement, juvenile court, and juvenile detention practitioners to monitor the pulse of their community; and consequently be in a more advantageous position to anticipate, and prepare for, any substantial operational and workload alterations which may be a product of changes in the community's social and economic environment.

Method

Data Collection

Data were collected from a variety of sources including published annual reports, administrative agency datasets, the Internet, and centralized data repositories (Refer to Tables 1 and 2 for a complete listing of the study variables and the sources from which they were compiled). All raw data were collected with the county being specified as the primary unit of analysis and were standardized to reflect either percentages or population rates. At the time of this study the most recently available juvenile crime data were from calendar year 1993. Consequently, this year was selected as the target data collection year for all study variables. In cases where 1993 calendar year data were unavailable data for the closest fiscal year were utilized. For variables enumerated in the United States Census the most recently collected data (1990) were used. Compared with juvenile arrest and delinquency petition filing rates training school admission rates are extremely low and demonstrate a substantially greater amount of annual variation both within, and between, counties. Consequently, the training school admission data were collected for a three year period in an effort to control, or minimize, the potential confounds associated with using only one year of data for variables displaying these statistical properties. Child homicide rates also exhibit this extreme variation and thus it was necessary to collect several years of data for this variable as well.

Data Analysis

The relatively large number of study variables and the possibility of multicollinearity, existing between many of the indicator variables, necessitated a two stage analytical process. Quartile rankings were created for the three dependent variables (juvenile arrest, delinquency petition filing, and training school admission rates) in order to facilitate the use of discriminant function analysis. A discriminant function analysis was conducted in order to identify those variables which were associated with changes in the juvenile arrest rate. This procedure was then repeated twice; once to identify those variables associated with the filing of delinquency petitions and again to identify the significant variables associated with the rate of training school admissions. Variables surviving these discriminant function analyses were retained and included in the second stage of data analysis. Variables that were not significantly associated with any of the juvenile crime measures were discarded. Stepwise multiple regression was used, in the second stage of data analysis, for narrowing the search to identify only the most salient and significant factors which demonstrated the greatest explanatory power for understanding why juvenile arrests, filings of delinquency petitions, and training school admissions vary from one county to another.

Table 1

Table 2

Results

Juvenile Arrests

Ten study variables were directly associated with juvenile arrest rates. Fifty percent of these factors were related to family structure and parenting style with juvenile arrest rates being significantly higher in those counties which possessed greater rates of divorce and single parent households. Higher rates of reported child abuse and neglect, a high rate of out of home placements, or children under the custody of social service agencies, and an excessive percentage of teenage births were more apparent in those counties with greater juvenile arrest rates. Juvenile arrest rates were significantly higher in more densely populated and urbanized counties, and in those counties reporting greater rates of children living in poverty. School dropout and child homicide rates were much more pronounced among counties with larger juvenile arrest rates. Combined, these ten factors are capable of statistically explaining 71% of the total variance in the counties' juvenile arrest rates.

Results from the stepwise regression indicate that two of these ten variables possess enough statistical power to account for nearly half of the total variance in juvenile arrest rates. Together the percentage of the population residing in urbanized areas and the rate of single parent households account for 41% of the variance. Counties with a greater number of single parent households, and a large urban population, tend to have significantly higher juvenile arrest rates when compared with counties with a larger percentage of dual parent urban, or rural, households and counties with a greater percentage of single parent rural households.

It should be noted that several of the most commonly held assumptions concerning juvenile arrests were not supported by this analysis. Contrary to prior research neither age, nor race, or gender exerted a significant main effect upon the counties' juvenile arrest rates. The arrest rates for counties with larger male, or African-American, or juvenile populations did not differ significantly from the arrest rates of counties with lower male, or African-American, or juvenile populations. Neither age, nor gender, nor racial composition can be postulated as a single, and sufficient, explanation for the differing arrest rates among the counties.

Delinquency Petition Case Filings

Eleven study variables displayed a statistically significant association with delinquency petition filing rates. Nine of these significant factors were related to the demographical and economic features of the county. In general, delinquency petition filing rates were substantially higher in those counties which tended to display a weak or declining economic infrastructure. Higher filing rates were more likely to be found among those counties with high unemployment rates, an excessive rate of AFDC recipients, a greater percentage of children receiving monthly AFDC support, and a high rate of children living in poverty. Larger delinquency petition filing rates were found to coincide with greater levels of mortality and violence within the county. Significantly higher filing rates existed for those counties demonstrating an excessive rate of reported homicides. Child mortality also exerted an influential effect on filing rates with significantly higher filing rates being found among those counties reporting higher child homicide and child death rates.

Other significant factors associated with the delinquency petition filing rate included: the number of single parent households per 1,000 households, the number of African-Americans as a percentage of the county population, and the percentage of the population residing in urban areas. Academic success, as reflected by the projected percentage of students expected to graduate, was negatively related to the filing rate; with lower filing rates being found in those counties with higher graduation estimates. These eleven factors were capable of explaining 67% of the total variance, or disparity, which exists between the counties' filing rates.

Results from the stepwise multiple regression reveal that of the eleven variables only two emerged as significant indicators of delinquency petition filing rates. As with the juvenile arrest rate percent of urbanization, and the rate of single parent households, serve as the most salient indicator or predictor variables when explaining why filing rates for delinquency petitions vary among the counties. However it should be noted that the explanatory or predictive power of these factors is marginal and extremely lower when compared to their power to explain varying juvenile arrest rates. The percent of the population residing in urban areas and the rate of single parent households only explained 17.5% of the total variance among the counties.

Training School Admissions

Many of the family and demographic attributes associated with juvenile arrest and delinquency petition filing rates were also significantly associated with the counties' training school admission rates. Training school admission rates were significantly higher under the same conditions previously discussed for delinquency petitions; however, unemployment and child death rates were not significantly related to training school admissions. As with juvenile arrests counties experiencing high divorce rates and a large percentage of teenage births were more likely to report higher training school admission rates. In addition; two variables, the percentage of babies born with low birth weights and the juvenile population as a percentage of the county population, were significantly and directly related to training school admission rates. When these percentages were excessive counties were more likely to report a greater rate of training school admissions. The variables associated with training school admission rates were capable of accounting for 72.5% of the total variance in these rates.

Consistent with the prior regression analyses the rate of single parent households, and the percentage of the population residing in urban areas, were identified as the strongest and most salient indicators of training school admission rates. These two family and demographic attributes explained 37% of the total variance in the counties' training school admission rates. As with juvenile arrest and delinquency petition filing rates counties with a high rate of single parent households, and a greater urban population, experienced substantially higher training school admission rates.

Results from the three separate discriminate function and multiple regression iterations failed to reveal any significant relationships among the three juvenile crime dependent measures and nine of the study variables. The number of males per 1,000 county residents, the percentage of net migration, the infant mortality rate, mean SAT scores, the number of school violence incidents per 1,000 students, median family income, and the per capita income did not demonstrate any significant and appreciable effects upon the counties' juvenile arrest, delinquency petition filing, or training school admission rates. The percentage of female headed households, with children, residing below the poverty level, as well as the poverty rate per 1,000 children, residing in households headed by a single female, also did not exert any significant effects upon the three juvenile crime variables. Consequently; these factors cannot be posited as primarily sufficient indicators, or causative sources, when explaining externally induced changes in the workload and case processing activities of the juvenile justice system and its major components. However, this finding does not preclude the ability of these variables to exert an interactive effect; an effect in which these variables operate in conjunction with other significant variables to either exacerbate, or mitigate against, increases in juvenile arrests, the filing of delinquency petitions, and training school admissions.

Discussion

While numerous factors were significantly associated with each of the three juvenile crime dependent measures only two of the study variables displayed this significant association across all three dependent measures. The research findings document a consistent and statistically significant link between the rate of single parent households and the percentage of the population residing in urban areas with the three major case processing activities of the juvenile justice system. While the explanatory power of these two factors varied, with each of the three dependent measures, they nonetheless possess enough statistical power to function as significant indicators, or predictors, of changing workload and case processing activities across the entire spectrum of the juvenile justice system. While a simultaneous increase in the rate of single parent households and the percentage of the population residing in urban areas cannot exactly predict, or guarantee with 100% accuracy, a corresponding increase in the juvenile arrest, delinquency petition filing, and training school admission rates it can serve as a more than adequate warning of possible potential increases in these rates. Monitoring these two factors will arm juvenile justice practitioners and policy makers with more than simple educated guesses, uninformed hunches, and random hit and miss approaches as they seek to address the juvenile crime problem and formulate effective solutions for this problem.

This research suggests a need for concentrating fiscal and programmatic resources, for prevention and intervention activities, within those areas of the community that possess a high rate of single parent households and a larger urban population. Prevention and intervention initiatives should target those children and teens residing in single parent households within the urban environment. The research findings further implicate the need for a multiple strategy approach with a multiple agency implementation plan. Community wide strategies and programs, which target these at risk children and teens at each critical contact point of the juvenile justice system, should be developed and implemented. The development and implementation of these comprehensive strategies and programs should entail extensive collaboration between law enforcement, juvenile court, and juvenile detention agencies and their personnel. These agencies should solicit contributions and advice from the existing community organizations and institutions, such as the school, church, medical, social services, and mental health facilities. As Hawkins, Catalano, and Brewer (1995) note the inclusion of these organizations will produce a more eclectic set of strategies and programs; strategies and programs which should be more efficacious in attacking not only juvenile crime, but also a broad range of related health and behavioral problems such as teen pregnancy, school dropout, and substance abuse.

From a policy perspective the importance of this study can be derived from the research goal of identifying the most salient and significant factors which exert an impact upon all of the major juvenile justice agencies and their respective caseload activities. While each component of the system is affected and influenced by its own unique set of family, school, community, and economic factors the identification of those influential factors common to juvenile law enforcement, courts, and detention will foster a systems approach to ameliorating juvenile crime. A unified or holistic approach, involving collaboration of all involved juvenile justice and community agencies, could alleviate system fragmentation, duplication of work, and cost overruns which often occur when each agency targets its own problems and formulates its own solutions independent of other agencies. Consequently; a comprehensive attack on a single set of underlying factors would be advantageous when a community's juvenile justice and juvenile service agencies face a scarcity of financial resources and\or limited operational abilities.

Future research should be conducted which examines why and how the rate of single parent households and a greater urban population interact to exert a significantly substantial impact upon the activities of North Carolina's juvenile justice system. The dynamics of the urban single parent household should be explored with an emphasis on isolating those family and community features that have the potential ability to alleviate, or at least mitigate, the negative effects associated with residing in the urban single parent home.

Replication of this study utilizing a smaller geographical domain, such as the city or preferably a single neighborhood, as the primary unit of analysis would be beneficial. The inclusion of at least ten years worth of time series data would also be beneficial and would contribute further understanding of those external factors which affect North Carolina's juvenile justice system and advance the knowledge of how these factors have historically shaped the system. Different family, school, community, demographic, and economic factors may emerge as statistically significant when different units of analyses are employed and when a longer temporal sequence is used for data collection. The inclusion of drug and weapon availability measures, which were unavailable for this initial exploratory study, should be incorporated into the research designs of future studies (Olsen, 1994, provides an excellent methodology for assessing and collecting data at the community or neighborhood level).

References

Blumstein, A. (1995). Violence by young people: Why the deadly nexus? National Institute ofJustice Journal, 229, 2-9.

Bursik, R.J., Jr., & Webb, J. (1982). Community change and patterns of delinquency. AmericanJournal of Sociology, 88, 24-42.

Butts, J.A., Snyder, H.N., Finnegan, T.A., Aughenbaugh, A.L., Tierney, N.J., Sullivan, D.P., & Poole, R.S. (1995). Juvenile court statistics 1992: Statistics report. Washington, D.C: Office of Juvenile Justice and Delinquency Prevention.

Elliott, D.S. (1993). Youth violence: An overview . Philadelphia: University of Pennsylvania, Center for the Study of Youth Policy.

Farrington, D. (1991). Childhood aggression and adult violence. In D. Pepler, & K.H. Rubin (Eds.), The development and treatment of childhood aggression (pp.5-29). Hillsdale, N.J: Lawrence Erlbaum.

Flanagan, T.J., & Maguire, K. (Eds.). (1990). Sourcebook of criminal justice statistics:1989. Washington, D.C: U.S. Department of Justice, Bureau of Justice Statistics.

Flanagan, T.J., & Maguire, K. (Eds.). (1992). Sourcebook of criminal justice statistics:1991. Washington, D.C: U.S. Department of Justice, Bureau of Justice Statistics.

Gottfredson, G.D. (1981). Schooling and delinquency. In S.E. Martin, L.B. Sechrest, & R. Redner (Eds.), New directions in the rehabilitation of criminal offenders. Washington, D.C: National Academy Press.

Hawkins, J.D., Catalano, R.F., & Brewer, D.D. (1995). Preventing serious, violent, and chronic juvenile offending: Effective strategies from conception to age 6. Unpublished manuscript.

Hawkins, J.D., Catalano, R.F., & Miller, J.Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychological Bulletin, 112 (1), 64-105.

Huizinga, D., Loeber, R., & Thornberry, T.P. (1994). Urban delinquency and substance abuse: Initial findings, research summary. Washington, D.C: Office of Juvenile Justice and Delinquency Prevention.

Kornhauser, R. (1978). Social sources of delinquency. Chicago: University of Chicago Press.

Loeber, R., Stouthamer-Loeber, M. (1986). Family factors as correlates and predictors of juvenile conduct problems and delinquency. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research (pp.29-149). Chicago: University of Chicago Press.

Loeber, R., Stouthamer-Loeber, M., Van Kammen, W., & Farrington, D.P. (1991). Initiation, escalation, and desistance in juvenile offending and their correlates. Journal of Criminal Law and Criminology, 82 (1) 36-82.

Maguire, K., & Pastore, A. (Eds.). (1994). Sourcebook of criminal justice statistics:1993. Washington, D.C: U.S. Department of Justice, Bureau of Justice Statistics.

Murray, C.A. (1983). The physical environment and community control of crime. In J. Q. Wilson (Ed.), Crime and public policy (pp.107-122). San Francisco: Institute for Contemporary Studies.

North Carolina Administrative Office of the Courts. (1988). North Carolina courts:1987-1988. Raleigh, N.C: North Carolina Administrative Office of the Courts.

North Carolina Administrative Office of the Courts. (1992). North Carolina courts:1991-1992. Raleigh, N.C: North Carolina Administrative Office of the Courts.

North Carolina Administrative Office of the Courts. (1994). North Carolina courts:1993-1994. Raleigh, N.C: North Carolina Administrative Office of the Courts.

North Carolina Child Advocacy Institute. (1995). Children's index: A profile of leading indicators of the health and well-being of North Carolina's children, 1995. Raleigh, N.C: Author.

North Carolina Department of Environment, Health, and Natural Resources, State Center for Health and Environmental Statistics. (1994). 1990-1993 child homicides and child deaths [ Machine readable data extract ]. Raleigh, N.C: State Center for Health and Environmental Statistics.

North Carolina Department of Human Resources, Division of Youth Services (1994). MIS summary: Fiscal year 1993-1994. Raleigh, N.C: Division of Youth Services.

North Carolina Department of Justice, State Bureau of Investigation. (1989). Crime in North Carolina: 1988 uniform crime report. Raleigh, N.C: Division of Criminal Information.

North Carolina Department of Justice , State Bureau of Investigation. (1993). Crime in North Carolina: 1992 uniform crime report. Raleigh, N.C: Division of Criminal Information.

North Carolina Department of Justice, State Bureau of Investigation. (1995). Crime in North Carolina: 1994 uniform crime report. Raleigh, N.C: Division of Criminal Information.

North Carolina Department of Public Instruction. (1995). Annual report on school violence: 1993- 1994 baseline year. Raleigh, N.C: North Carolina Department of Public Instruction.

North Carolina Office of State Planning, State Data Center. (1991). 1990 LINC census data [ Machine readable data file ]. Raleigh, N.C: State Data Center.

North Carolina Office of State Planning, State Data Center. (1993). 1992 LINC data [ Machine readable data file ]. Raleigh, N.C: State Data Center.

Olsen, J. (1994). Preventing juvenile violence and delinquency, risk focused prevention using communities that care: Risk and resource assessment training from Developmental Research and Programs, Inc. Seattle: Developmental Research and Programs, Inc.

Pope, C., & Feyerherm, W. (1991). Minorities and the juvenile justice system. Washington, D.C: Office of Juvenile Justice and Delinquency Prevention.

Snyder, H., & Sickmund, M. (1995). Juvenile offenders and victims: A focus on violence, statistics summary. Washington, D.C: Office of Juvenile Justice and Delinquency Prevention.

Snyder, H., & Sickmund, M. (1995). Juvenile offenders and victims: A national report. Washington, D.C: Office of Juvenile Justice and Delinquency Prevention.

Wilson, J.Q., & Herrnstein, R.J. (1985). Crime and human nature. New York: Simon & Schuster.

Wright, K.N., & Wright, K.E. (1994). Family life, delinquency, and crime: A policymaker's guide. Washington, D.C: Office of Juvenile Justice and Delinquency Prevention