Personal Criminal Victimization in the United States: Fixed and Random Effects of Individual and Household Characteristics

The present research is concerned with multilevel modeling of personalvictimization rates. Data from the 1994 National Crime Victimization Surveyare employed. Following the routine activities and lifestyle victimizationtheory, individuals' profile and lifestyle as well as characteristics oftheir household comprise the set of explanatory variables. The results ofestimated multilevel negative binomial models, which explicitly disentanglethe unexplained heterogeneity between individuals and between households,are discussed. The estimated random effects of household characteristicsshow that the unexplained heterogeneity for the average number of personalcrimes differs across household types. Further, the individual covariateswith between households random effects become less influential the more thebase personal crime rates are high.