This study demonstrates that the empirical literature on the structural covariates of homicide rates contains inconsistent findings across different time periods and different geographical units. This apparent variance of findings may be due to statistical or methodological artifacts of particular studies, such as different time periods covered, units of analysis, samples, model specifications, and problems of statistical analysis and inference. A baseline regression model using 11 structural covariates is estimated for cities, metropolitan areas, and states in 1960, 1970, and 1980. The empirical estimates of this model exhibit instability because of high levels of collinearity among several regressors. Principal components analysis is applied to simplify the dimensionality of the structural covariate space. Reestimation of the regression model then indicates that the apparent inconsistencies across time and social space are greatly reduced. The theoretical significance of the findings for substantive theories of violent crime are discussed. The past two decades have seen an upsurge in the number of empirical studies that address the question of why some cities, metropolitan areas, ' The research reported here was supported, in part, by National Science Foundation grant SES-8800595. We are grateful to five anonymous A J S referees for stimulating and useful comments on an earlier version of this paper. Requests for reprints should be sent to

This article reviews questions about different categories of criminal careers, summarizes Poisson Iatent class regression models, describes procedures for evaluating the optimal number of latent classes, and applies this methodology to data from male cohorts taken from the cities of London, Philadelphia, and Racine. Four latent classes of offending careers is an appropriate number for the London cohort, but five classes can be justified for the Philadelphia data. In the case of the Racine cohorts, five classes may be detected for the 1942 and 1955 cohorts but only four for the 1949 cohort. Despite the varying numbers of latent offending classes, there clearly is a small number of typical age patterns. The Poisson distribution and Poisson regression models are fast becoming key methodological tools for the development of micromodels of delinquentlcriminal careers in the criminological literature (

Specifications and moment properties of the univariate Poisson and negative binomial distributions are briefly reviewed and illustrated. Properties and limitations of the corresponding poisson and negative binomial (gamma mixtures of Poissons) regression models are described. It is shown how a misspecification of the mixing distribution of a mixed Poisson model to accommodate hidden heterogeneity ascribable to unobserved variables—although not affecting the consistency of maximum likelihood estimators of the Poisson mean rate parameter or its regression parameterization—can lead to inflated t ratios of regression coefficients and associated incorrect inferences. Then the recently developed semiparametric maximum likelihood estimator for regression models composed of arbitrary mixtures of Poisson processes is specified and further developed. It is concluded that the semiparametric mixed Poisson regression model adds considerable flexibility to Poisson-family regression models and provides opportunities for interpretation of empirical patterns not available in the conventional approaches.

This research examines the differential effects of structural conditions on race-specific victim and offender homicide rates in large U S . cities in 1990. While structural theories of race relations and criminological explanations are reviewed, particular attention is given to those structural theories that highlight racial competition, economic and labor market opportunity, and racial segregation as essential for an examination of racially disaggregated homicide offending. The effects of these and other structural conditions are estimated for four racially distinct homicide offending models-black intraracial, white intraracial, black interracial, and white interracial homicides. The results suggest that the structural conditions that lead to race-specific victim and offender homicide rates differ significantly among the four models. Economic deprivation and local opportunity structures are found to influence significantly the rates of intraracial homicide offending, while racial inequality contributes solely to black interracial homicide rates. In addition, our findings indicate that blacks and whites face different economic and social realities related to economic deprivation and social isolation. The differential impact of these structural conditions and other labor market factors are discussed.Criminologists long have studied the question of what differences exist between blacks and whites in their involvement in homicide. Recently, researchers have provided some answers to this question when examining separate black and white homicide offending models (Harer and Steffensmeier

As the 20-year mark since the publication of an article by Kenneth C. Land, Patricia L. McCall, and Lawrence Cohen, “Structural Covariates of Homicide Rates: Are There Any Invariances Across Time and Social Space?” approaches, the question that these scholars originally posed is raised again: Have researchers been able to identify a set of robust structural covariates that consistently predict crime rates? Subsequent to the publication of this piece, numerous scholars have replicated and extended its conceptual, methodological, and empirical work in various ways—with more than 500 citations to date. In response to this attention, the authors first review the advances made by the Land et al. article. This is followed by a review of findings from studies published over the past 20 years to determine which structural predictors identified in the Land et al. piece continue to be prominent in the study of homicide and which structural predictors have surfaced in recent years as influential to crime rates. Using data on U.S. cities for the years 1970, 1980, 1990, and 2000, the authors then present a systematic empirical assessment of the explanatory power of the covariates of homicide rates identified in the Land et al. study. Twenty years later, we find support for the claims of invariance established in Land et al. and acknowledge the contributions of this piece to the macrolevel study of homicide rates.