Economic Disadvantage and Homicide

• Author: Karen F. Parker,Richard Stansfield,Kirk R. Williams
• DOI: 10.1177/1088767916647990
• Date: 01 February 2017
• Published date: 01 February 2017
• Subject Matter: Articles

Homicide Studies

2017, Vol. 21(1) 59 –81

© 2016 SAGE Publications

Reprints and permissions:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1088767916647990

journals.sagepub.com/home/hsx

Article

Economic Disadvantage

and Homicide: Estimating

Temporal Trends in

Adolescence and Adulthood

Richard Stansfield1, Kirk R. Williams2,

and Karen F. Parker2

Abstract

Although research has established economic disadvantage as one of the strongest,

most robust predictors of urban violence, the conditions under which this relation

holds need further elaboration. This study examines the disadvantage–violence link

across age-specific transitional periods from adolescence to adulthood and provides

theoretical arguments for why the strength of this relation should decline with age.

Using 90 of the largest cities in the United States, the present study analyzes the

impact of economic disadvantage and other urban conditions (residential instability,

family disruption, and population heterogeneity) on age-specific homicide counts

from 1984 to 2006. The analytical strategy incorporates temporal trends by using

negative binomial fixed-effects regression models. The results reveal a consistent

decline from adolescence to adulthood in the strength of the estimated effects of

economic disadvantage, residential instability, and family disruption on homicide

trends. The findings are discussed in terms of the implications for future research

and public policy.

Keywords

structural causes, age, correlates, social disorganization, economic disadvantage

In their landmark study assessing covariates of homicide rates, Land, McCall, and

Cohen (1990) examined the use of structural covariates of homicide in 30 years of

1Rutgers University, Camden, NJ, USA

2University of Delaware, Newark, USA

Corresponding Author:

Richard Stansfield, Rutgers University, 405-407 Cooper Street, Camden, NJ 08102, USA.

Email: rstans@udel.edu

647990HSXXXX10.1177/1088767916647990Homicide StudiesStanseld et al.

research-article2016

60 Homicide Studies 21(1)

criminology literature, showing considerable inconsistency in the empirical support

for a number of these covariates. Proposing avenues to overcome inconsistencies

using methodological strategies, they illustrated how a multidimensional measure, like

the economic disadvantage index, produces consistent and strong results across time

and multiple levels of analysis. Indeed, scholars have noted that “multiple forms of

disadvantage . . . tend to come bundled together in urban neighborhoods” (Friedson &

Sharkey, 2015, p. 342); thus, finding more consistent results with a measure that cap-

tures those multiple forms should be expected. Accordingly, over time, scholars have

shown disadvantage to be a robust determinant of homicide within and across urban

areas in the United States (Peterson & Krivo, 2005; Pratt & Cullen, 2005; Pridemore,

2002).

Peterson and Krivo (2010) provided a striking example. They analyzed sociodemo-

graphic data from the U.S. Census and crime data reported to the police in 2000. Using

data that covered almost 9,000 urban neighborhoods within 87 cities in the United

States, their baseline model indicated that neighborhood violence rates for African

Americans were 327% higher than for Whites. Adding their multidimensional mea-

sure of disadvantage to the model reduced that differential to 65%. No other neighbor-

hood characteristic had anywhere near the substantial impact of disadvantage.

The research by Peterson and Krivo (2010), like most of its predecessors, focused

on variation in violence rates across geographic units such as neighborhoods, census

tracts, or cities (see Kovandzic, Vieratis, & Yeisley, 1998; Parker & McCall, 1999, for

reviews). Other investigators explored the disadvantage–violence link using temporal

(McCall, Parker, & MacDonald, 2008; McDowell & Loftin, 2009; Ousey & Lee,

2002) and latent trajectory analytic approaches (McCall, Land, & Parker, 2011), all

finding support for this relation. Further research is needed to understand the relation

between disadvantage and violence. For example, most of those analyses did not dis-

aggregate data by age to determine whether disadvantage and other macro-level

covariates have differential effects on homicide rate variation over time and across age

groups.

Studies addressing differential effects by age are present in the literature, but they

have limitations. For example, some research estimated effects of disadvantage on

homicide, but the analysis was limited to a single age group, was cross-sectional in

design, and focused on a specific type of homicide (e.g., Ousey & Augustine, 2001).

Other research was longitudinal but utilized only two periods, with comparisons lim-

ited to two age groups (e.g., MacDonald & Gover, 2005; Strom & MacDonald, 2007).

As an example, studies found that a rise in indicators of economic disadvantage was

associated with a rise in city-level rates of youth-on-youth homicide (MacDonald &

Gover, 2005), and youth homicide perpetrated by White and Black youth (Strom &

MacDonald, 2007). Baumer (2008) conducted a comprehensive time-series analysis

of homicide and other conventional street crimes, comparing youth (ages 15-24) and

adults (ages 25-44), but he did not include a multidimensional measure of disadvan-

tage. He used separate economic indicators: job availability, unemployment rate, and

county average real wages, with only the latter measure having a significant negative

estimated effect on youth but not adult homicide. Despite the limitations of previous