Ecological analysis of crime rates and police discretion with young persons: a replication.

• Author: Schulenberg, Jennifer L.
• Position: Canada

Previous research published in this journal used 1991 data to conduct an ecological analysis of the police use of formal social control (see Schulenberg 2003). This note replicates the analysis using 2001 data on municipal crime rates and police discretion with apprehended youth in Canada. We also extend the earlier research by including additional indicators of social disorganization and of criminal opportunity that were not previously included in the analysis.

Background

Schulenberg (2003) explored the applicability of four ecological theories of crime--urbanization, social disorganization, opportunity theory, and the overload hypothesis--to Canadian municipal crime rates and the use by police of formal social control with apprehended youth. Data for 1991 from the Uniform Crime Reporting (UCR) Survey (Canadian Centre for Justice Statistics 2002), the Police Administration Survey (Canadian Centre for Justice Statistics 2001), and the census (Statistics Canada 2002b) (2) were used to assess the applicability of the four theories within an integrated causal model, by regressing indicators of the crime rate and of police charging practices on the relevant attributes of police forces and municipalities. Only social disorganization theory was supported as an explanation of the municipal crime rate. Urbanization theory and social disorganization theory were supported as explanations of police charging practices with apprehended youth.

Other research based on these theories has used indicators that were not available to Schulenberg's research: for example, ethnic heterogeneity as an indicator of social disorganization and the employment status of married women as an indicator of criminal opportunity (Schulenberg 2003: 150). The present research replicates Schulenberg (2003) by employing the same causal model (see Figure 1), similar data sources, and the same methods of analysis. However, the data are more recent (2001 versus 1991), and additional indicators of key concepts are included. The reader is referred to Schulenberg (2003: 129-137) for an exposition of the four theories and the rationale for the integrated causal model shown in Figure 1.

[FIGURE 1 OMITTED]

Data

Dependent variables

While there are several endogenous variables in the causal model (Figure 1), the two dependent variables of theoretical interest are the municipal crime rate and the use by police of formal social control with apprehended youth. Following Schulenberg (2003), the crime rate was operationalized in the usual way by the number of actual criminal incidents reported by police per 100,000 municipal population, taken from the UCR (Canadian Centre for Justice Statistics 2002). Unlike Schulenberg (2003), this was logged in order to reduce skew and outliers. Descriptive statistics for all variables are given in Table 1.

As in Schulenberg (2003), the use by police of formal social control with youth was operationalized by the "youth charge ratio," also calculated from UCR data (Canadian Centre for Justice Statistics 2002). The youth charge ratio is the number of young persons charged divided by the numbers of young persons apprehended (which, in turn, is the sum

of young persons charged and young persons who could have been charged but were dealt with informally) (Schulenberg, 2003).

Both dependent variables included all federal statute offences and excluded provincial and municipal offences. While some research at the level of the individual suggests there are distinctions in the factors related to different types of crimes (i.e., property crimes vs. violent crimes), this is not common in ecological research (Schulenberg 2003). In a recent ecological study, Ouimet (2000: 149-150) compared the factors related to violent and non-violent crime in Montreal and concluded that the impacts of these factors were similar for both types of crime. Furthermore, there is no theory or research suggesting that ecological factors influence police differently when they are responding to different types of crime (Schulenberg 2003). Therefore, following Schulenberg (2003), the present analysis does not distinguish among types of crime.

Independent variables

Data for the independent variables come from two sources: the Census (Statistics Canada 2002b) (for community characteristics), and the Police Administration Survey (for indicators of police workload) (Canadian Centre for Justice Statistics 2001). Although data for the UCR (Canadian Centre for Justice Statistics 2002b) and the Police Administration Survey are collected annually, Census data are collected every five years, so 2001 represents the most recent data available.

As in Schulenberg (2003), the indicator of urbanization in the present analyses is the logged municipal population. Schulenberg (2003) used four indicators of social disorganization: the unemployment rate, the mover rate (the percentage of municipal residents who had moved in the past 12 months), the rental rate (the percentage of dwellings that were rented rather than owner-occupied), and the lone-parent family rate (the percentage of families headed by a lone parent). The present research also uses these indicators and adds three more: high occupational status (the percentage of adults in high-status occupations), high education (the percentage of adults with post-secondary education), and ethnic heterogeneity. Following Sampson and Groves (1989: 784), the index of ethnic heterogeneity is defined as,

1 - ([summation] [p.sup.2.sub.i])

where [p.sub.i] is the fraction of the population in a given group.

The original Census categories (Statistics Canada 2002a: 74) were aggregated into the following groups, following the Statistics Canada 2001 Census Dictionary (Statistics Canada, 2003: Appendix C): White (non-minority), South/East/Southeast Asians, Arab/ West Asians, Blacks and Latin Americans, and other or mixed minorities.

One limitation of the research reported in Schulenberg (2003) was the use of the unemployment rate both as an indicator of social disorganization and as the sole indicator of criminal opportunity. The present research is able to assess the applicability of criminal opportunity theory more adequately by using the unemployment rate only as an indicator of social disorganization and using four other indicators of criminal opportunity. These variables capture the extent to...