The general purpose of this project was to examine homicide trends in Chicago neighborhoods from 1980-2000. Specifically, this study sought to:
- Assess neighborhood variation in homicide trends.
- Identify the particular types of homicide trajectory that Chicago neighborhoods follow.
- Assess whether structural characteristics of neighborhoods influence homicide trends and trajectories.
- Determine the extent to which the influence of structural characteristics are mediated by neighborhood levels of collective efficacy.
Data for this project was compiled from a variety of sources. To generate the homicide rate (per 100,000), homicide incident counts were obtained from HOMICIDES IN CHICAGO, 1965-1995 (ICPSR 6399). This dataset includes tract identifiers that allow for merging with other tract characteristics. The population sizes used in the denominator of the homicide rate were obtained from the decennial United States Census files for each decade from 1980-2000, and cubic spline interpolation was used to generate intercensal estimates. The homicide rate was calculated per 100,000 population. Due to a high degree of skewness in the homicide rate, the study analyses used the natural log of the homicide rate. One objective was to examine the relationships between homicide trends and various characteristics of neighborhood structure identified as relevant by the social disorganization perspective. These variables were drawn from the CENSUS OF POPULATION AND HOUSING, 1980 [UNITED STATES]: SUMMARY TAPE FILE 3A (ICPSR 8071), CENSUS OF POPULATION AND HOUSING, 1990 [UNITED STATES]: SUMMARY TAPE FILE 3A (ICPSR 9782), and CENSUS OF POPULATION AND HOUSING, 2000 [UNITED STATES]: SUMMARY FILE 3, NATIONAL (ICPSR 13396). In order to avoid the problems associated with high levels of multicollinearity among the regressors, several standardized mean indexes were created to represent key concepts identified in previous empirical and theoretical work. All predictors were incorporated as both static levels at the beginning of the time period, and as change scores for the period 1980 to 2000.
As an initial attempt at determining whether the effect of these structural characteristics on neighborhood crime trends are mediated by emergent properties of neighborhoods such as social cohesion, collective efficacy, and disorder, this study also incorporated data from the PROJECT ON HUMAN DEVELOPMENT IN CHICAGO NEIGHBORHOODS: COMMUNITY SURVEY, 1994-1995 (ICPSR 2766). This survey contained responses from interviews conducted with 8,782 Chicago residents comprising a representative sample for each of the 343 neighborhood clusters of Chicago. The resulting dataset includes measures such as willingness to engage in informal social control, social cohesion and trust, and social and physical disorder. The community survey was conducted in 1994-1995, so the timing was not ideal for incorporation into a study of homicide trends from 1980-2000. Therefore, this portion of the study analyses was restricted to the years 1990-2000.
Census tracts were excluded from the study if they had a population of less than 100 persons in any given year because such small denominators result in considerable instability in homicide rates. An additional two tracts (1701 and 3204) were omitted because they have no households. After removing these census tracts, the sample size was 827 tracts.
This study incorporated data from the PROJECT ON HUMAN DEVELOPMENT IN CHICAGO NEIGHBORHOODS: COMMUNITY SURVEY, 1994-1995 (ICPSR 2766). This survey contained responses from interviews conducted with 8,782 Chicago residents comprising a representative sample for each of the 343 neighborhood clusters of Chicago.
All census tracts in Chicago between 1980 and 2000.
HOMICIDES IN CHICAGO, 1965-1995 (ICPSR 6399)
PROJECT ON HUMAN DEVELOPMENT IN CHICAGO NEIGHBORHOODS: COMMUNITY SURVEY, 1994-1995 (ICPSR 2766)
CENSUS OF POPULATION AND HOUSING, 2000 [UNITED STATES]: SUMMARY FILE 3, NATIONAL (ICPSR 13396)
CENSUS OF POPULATION AND HOUSING, 1990 [UNITED STATES]: SUMMARY TAPE FILE 3A (ICPSR 9782)
CENSUS OF POPULATION AND HOUSING, 1980 [UNITED STATES]: SUMMARY TAPE FILE 3A (ICPSR 8071)
program source code
Variables used in the study analyses include census tract, natural log of the homicide rate, neighborhood characteristics, several standardized mean indexes, and community-level measures. Neighborhood characteristics variables derived from Census data include population size, median household income, poverty rate, unemployment rate, divorce rate, percentage of female-headed households, residential mobility, homeownership rate, racial and Hispanic composition, immigrant composition, percentage of households speaking English, and population density. The standardized mean indexes variables are concentrated disadvantage, family disruption, social disorganization, and immigrant concentration. The study also utilized a neighborhood ties measure and a disorder index.
In order to avoid the problems associated with high levels of multicollinearity among the regressors, several standardized mean indexes were created. The concentrated disadvantage index is comprised of median household income, the percentage of persons with a high school diploma, the percentage of persons with a bachelor's degree, percentage of persons who are African American, and the percentage of persons unemployed. Family disruption is an index comprised of the percentage of persons age 15 and over who are divorced and the percentage of children who live with a single parent. Social disorganization is measured with an index comprised of the percentage of persons who lived in a different house five years earlier, the percentage of homes that are renter-occupied, the percentage of housing units that are vacant, and population density. Immigrant concentration is measured using an index comprised of the percentage foreign-born and the percentage Hispanic.
The measure of neighborhood ties is an mean index derived from five survey items asking how often people in the neighborhood do favors for one another, how often neighbors watch over each other's property, how often neighbors ask for advice, how often people in the neighborhood get together, and how often people in the neighborhood visit in each other's homes. The disorder index is comprised of items measuring perceived levels of both social and physical disorder, including questions asking how much of a problem is posed by litter, graffiti, vacant buildings, drinking in public, using or selling drugs in public, and groups of teenagers or adults hanging out and causing trouble.