Age Structure and Neighborhood Homicide: Testing and Extending the Differential Institutional Engagement Hypothesis

• Author: Patricia L. McCall,Cindy Brooks Dollar,Joshua Fink,Kenneth C. Land
• Published date: 01 November 2017
• Date: 01 November 2017
• DOI: 10.1177/1088767917702474
• Subject Matter: Articles

https://doi.org/10.1177/1088767917702474

Homicide Studies

2017, Vol. 21(4) 243 –266

© 2017 SAGE Publications

Reprints and permissions:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1088767917702474

journals.sagepub.com/home/hsx

Article

Age Structure and

Neighborhood Homicide:

Testing and Extending the

Differential Institutional

Engagement Hypothesis

Cindy Brooks Dollar1, Patricia L. McCall2,

Kenneth C. Land3, and Joshua Fink3

Abstract

We examine the empirical applicability of differential institutional engagement in

explaining the youth age structure effect on neighborhood homicide. Using the

National Neighborhood Crime Study and Census data, we conduct a multilevel

spatial analysis of homicides in 8,307 census tracts. We find support for three

indicators of differential institutional engagement (disengaged youth, educational

engagement, employment engagement). An additional dimension of institutional

engagement (familial engagement) operates in the expected direction but is not

statistically significant. We argue that previous cross-sectional studies reporting a

null or negative relationship between percentage of young and homicide are due to

omitting measures of institutional youth (dis)engagement.

Keywords

homicide, age structure, neighborhood, institutional engagement, spatial analysis,

multilevel modeling

Criminologists have a rich history in examining the link between age and crime.

Dating back hundreds of years, social scientists have been involved in identifying

patterns of age structure and delinquency (e.g., Durkheim, 1897/1951; Quetelet,

1University of North Carolina at Greensboro, USA

2North Carolina State University, Raleigh, USA

3Duke University, Durham, NC, USA

Corresponding Author:

Cindy Brooks Dollar, Department of Sociology, University of North Carolina at Greensboro, 337 Frank

Porter Graham Building, P.O. Box 26170, Greensboro, NC 27402-6170, USA.

Email: cbdollar@uncg.edu

702474HSXXXX10.1177/1088767917702474Homicide StudiesDollar et al.

research-article2017

244 Homicide Studies 21(4)

1831/1984). Interest in age structure effects on crime rates partially stems from indi-

vidual-level studies of criminal offending. Although some divergences in individual-

level patterns of criminal offending have been identified, researchers generally

conclude that both participation in, and victimization by, crime increases from the

mid-teens to late 20s and then begins to steadily decline over the life course. Abstracting

these individual-level findings to population aggregates of various levels (e.g., states,

metropolitan areas, cities, and neighborhoods), a positive relationship between the

relative size of youth populations and rates of crime is expected. However, researchers

have produced a vast body of inconsistent findings for the youth age structure–crime

rate relationship often citing negative or null relationships between the two (Marvell

& Moody, 1991; Pratt & Cullen, 2005).

Although various theoretical frameworks have been used to explain the association

between youth age structure and crime, scholars have recently begun focusing on cer-

tain ecological contingencies to explicate varying age structure effects on crime (e.g.,

Phillips, 2006; Thomas & Shihadeh, 2013). In their recent article, McCall, Land,

Dollar, and Parker (2013) proposed the concept of differential institutional engage-

ment to explain the incongruous findings of the youth age structure–crime relation-

ship.1 Contending that youth populations are heterogeneous in their level of institutional

attachment and relying on theoretical arguments and empirical evidence suggesting

that engagement in prosocial institutions deters criminal involvement, McCall et al.

(2013) hypothesized that the link between age structure and crime rates would become

consistent and operate in the expected direction if the youth population’s relative

involvement was considered and specified. Their city-level analysis of homicide rates

in 1980, 1990, and 2000 provided support for their postulations and purports to move

us forward in resolving the age–crime puzzle.

The notion of differential institutional engagement relies heavily on social control

mechanisms theorized to occur through neighborhood or community relations, but the

question remains as to whether or not differential institutional engagement can explain

spatial variation in crime rates in more localized units. The goal of the present article

is twofold. First, we elaborate the concept of differential institutional engagement by

examining additional measures of the construct. Second, we assess the ability of this

construct to explain the youth age structure–crime relationship (as measured by the

homicide rate) at a lower level of aggregation (neighborhoods, as measured by census

tracts) than analyzed in McCall et al. (2013). In short, we examine whether the rela-

tionship of neighborhood youth age structure to homicide becomes consistently posi-

tive when youth institutional (dis)engagement is controlled. In doing so, we are

suggesting that previous homicide studies reporting null or negative relationships

between percentage of young and homicide may be due to omitted variable bias (i.e.,

neglecting measures of institutional youth (dis)engagement).

Inconsistencies of Youth Age Structure Effects in

Ecological Studies of Crime

Criminologists generally agree that criminal offending and victimization peaks from

mid-teens to young adulthood after which it decreases with age across the life course.