Journal of Quantitative Criminology 27(4)

Journal of Quantitative Criminology, December 2011: Volume 27, Issue 4

Static and Dynamic Indicators of Minority Threat in Sentencing Outcomes: A Multi-Level Analysis
Cyndy Caravelis, Ted Chiricos and William Bales

Designation as a “Habitual Offender” is an enhanced form of punishment which unlike, “Three Strikes” or “10-20-Life,” is entirely discretionary. We use Hierarchical Generalized Linear Modeling to assess the direct effects of race and Latino ethnicity on the designation of Habitual Offenders as well as the effect of both static and dynamic indicators of racial and ethnic threat on those outcomes. Our data include 26,740 adults sentenced to prison in Florida between 2002 and 2004 who were statutorily eligible to be sentenced as Habitual. The odds of receiving this designation are significantly increased for black and Latino defendants as compared to whites, though race and ethnicity effects vary substantially by crime type, being strongest for drug offenses and negligible for violent crimes. Static measures of group level threat (% black and % Latino) have no cross-level effect on sentencing by race or Latino ethnicity. However, increasing black population over time increases the odds of being sentenced as Habitual for both black and Latino defendants. Increasing Latino population increases the odds of Habitual Offender sentencing for Latinos, but decreases it for blacks. The prospect of engaging dynamic as opposed to static measures of threat in future criminal justice and other social control research is discussed.

Examining the Neighborhood Context of the Violent Offending-Victimization Relationship: A Prospective Investigation
Mark T. Berg and Rolf Loeber

The persistent link between offending and victimization is one of the most robust empirical findings in criminological research. Despite important efforts to isolate the sources of this phenomenon, it is not fully understood. Much attention has been paid to the role of individual-level factors; however, few studies have systematically integrated neighborhood conditions. Using prospective data from the Pittsburgh Youth Study the current research examines a set of hypotheses regarding the interplay of neighborhood structural conditions and the victim-offender overlap. A multilevel analytical technique is applied to the data which purges time-varying covariates of all time-stable unobserved heterogeneity. Results indicate that the relationship between offending and victimization is pronounced in disadvantaged neighborhoods, while offending is not significantly related to victimization risk in contexts marked by lower levels of disadvantage. The implications of the results for theory are discussed, along with recommendations for future research.

Racial Disparity in Police Stop and Searches in England and Wales
Vani K. Borooah

Data published by the United Kingdom’s Ministry for Justice clearly shows that, compared to persons who were White, members of racial minorities in England, particularly Blacks, were far more likely to be stopped and searched by the police. The question is whether such racial disparity in stops and searches could be justified by racial disparities in offending? Or whether the disparity in stop and searches exceeded the disparity in offending? This paper proposes a method for measuring the amount of excess in racial disparity in police stop and searches. Using the most recently published Ministry of Justice data (for 2007/08) for Police Areas in England and Wales it concludes that while in several Areas there was no excess to racial disparity in police stop and searches, there was, on the basis of the methodology proposed in the paper, evidence of such excess in some Police Areas of England and Wales.

Structural Determinants of Homicide: The Big Three
Maria Tcherni

Building upon and expanding the previous research into structural determinants of homicide, particularly the work of Land, McCall, and Cohen (1990), the current paper introduces a multilevel theoretical framework that outlines the influences of three major structural forces on homicidal violence. The Big Three are poverty/low education, racial composition, and the disruption of family structure. These three factors exert their effects on violence at the following levels: neighborhood/community level, family/social interpersonal level, and individual level. It is shown algebraically how individual-level and aggregate-level effects contribute to the size of regression coefficients in aggregate-level analyses. In the empirical part of the study, the presented theoretical model is tested using county-level data to estimate separate effects of each of the Big Three factors on homicide at two time periods: 1950–1960 and 1995–2005 (chosen to be as far removed from one another as the availability of data allows). All major variables typically used in homicide research are included as statistical controls. The results of analyses show that the effects of the three major structural forces—poverty/low education, race, and divorce rates—on homicide rates in US counties are remarkably strong. Moreover, the effect sizes of each of the Big Three are found to be identical for both time periods despite profound changes in the economic and social situation in the United States over the past half-century. This remarkable stability in the effect sizes implies the stability of homicidal violence in response to certain structural conditions.

Estimating the Impact of Classification Error on the “Statistical Accuracy” of Uniform Crime Reports
James J. Nolan, Stephen M. Haas and Jessica S. Napier

This paper offers a methodological approach for estimating classification error in police records then determining the statistical accuracy of official crime statistics reported to the Uniform Crime Reporting (UCR) program. Classification error refers to the mistakes in UCR statistics caused by the misclassification of criminal offenses, for example recording a crime as aggravated assault when it should have been simple assault. Statistical accuracy refers to the estimated true total of each crime type based on cancelling effect of undercounting and overcounting crime due to misclassifications. The population for the study consists of the 12 largest municipal police agencies in a mostly rural southeastern state. Based on a sample of 2,663 records, the authors illustrate the impact of classification error on the total population of reported offenses. Misclassifications result in overcounting and undercounting certain crimes. The true number of each crime type, as well as the aggregate Index Crime, Violent Crime, and Property Crime totals, is estimated based the evaluation of offsetting misclassifications. The findings show that certain UCR crime categories are greatly undercounted while others are overcounted. The index crime and violent crime totals are also significantly undercounted; however, when simple assault is added to the index and violent crime categories, the error in these aggregate numbers is reduced to less than 1%. The results provide a benchmark for assessing the statistical accuracy of the UCR data.

Spatializing the Social Networks of Gangs to Explore Patterns of Violence
George E. Tita and Steven M. Radil

The majority of spatial studies of crime employ an inductive approach in both the modeling and interpretation of the mechanisms of influence thought to be responsible for the patterning of crime in space and time. In such studies, the spatial weights matrix is specified without regard to the theorized mechanisms of influence between the units of analysis. Recently, a more deductive approach has begun to gain traction in which the theory of influence is used to model influence in geographic space. Using data from Los Angeles, we model the spatial distribution of gang violence by considering both the relative location of the gangs in space while simultaneously capturing their position within an enmity network of gang rivalries. We find that the spatial distribution of gang violence is more strongly associated with the socio-spatial dimensions of gang rivalries than it is with adjacency-based measures of spatial autocorrelation.

A Comparison of Logistic Regression, Classification and Regression Tree, and Neural Networks Models in Predicting Violent Re-Offending
Yuan Y. Liu, Min Yang, Malcolm Ramsay, Xiao S. Li and Jeremy W. Coid

Previous studies that have compared logistic regression (LR), classification and regression tree (CART), and neural networks (NNs) models for their predictive validity have shown inconsistent results in demonstrating superiority of any one model. The three models were tested in a prospective sample of 1225 UK male prisoners followed up for a mean of 3.31 years after release. Items in a widely-used risk assessment instrument (the Historical, Clinical, Risk Management-20, or HCR-20) were used as predictors and violent reconvictions as outcome. Multi-validation procedure was used to reduce sampling error in reporting the predictive accuracy. The low base rate was controlled by using different measures in the three models to minimize prediction error and achieve a more balanced classification. Overall accuracy of the three models varied between 0.59 and 0.67, with an overall AUC range of 0.65–0.72. Although the performance of NNs was slightly better than that of LR and CART models, it did not demonstrate a significant improvement.