The Relationship between Job Quality and Crime: Examining Heterogeneity in Treatment and Treatment Effect

• Date: 01 November 2020
• Author: Chae M. Jaynes
• DOI: 10.1177/0022427820918899
• Published date: 01 November 2020
• Subject Matter: Articles

Article

The Relationship

between Job Quality

and Crime: Examining

Heterogeneity in

Treatment and

Treatment Effect

Chae M. Jaynes

1

Abstract

Objectives: This study evaluates the relationship between employment and

crime through a holistic evaluation of both treatment and treatment effect

heterogeneity. Methods: This study implements a perceptual measure of job

quality (job satisfaction) and hybrid fixed effects models among a sample of

high-risk adults. Analyses also consider the robustness of findings across

alternative operationalizations of job quality and various sample subgroups.

Results: Transitioning from not working to working in the lowest quality job

can be criminogenic. Among those who are working, an improvement in job

quality is not generall y associated with offen ding. However, model an d

crime-specific effects are observed. Evidence of treatment effect hetero-

geneity is also found, suggesting the effect of job quality is moderated by

race/ethnicity and location. Conclusions: These findings caution criminolo-

gists against making an assumption that employment is inversely related to

1

Department of Criminology, University of South Florida, Tampa, FL, USA

Corresponding Author:

Chae M. Jaynes, Department of Criminology, University of South Florida, 4202 E. Fowler Ave.,

SOC 332, Tampa, FL 33620, USA.

Email: jaynes@usf.edu

Journal of Research in Crime and

Delinquency

2020, Vol. 57(6) 693-740

ªThe Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0022427820918899

journals.sagepub.com/home/jrc

offending and call into question our understanding of job quality as a general

disincentive for crime. Rather, evidence suggests that improvements in job

quality may result in modest reductions in offending, but only for certain

types of crime and certain individuals within specific labor market contexts.

Keywords

rational choice theory, criminological theory, employment, job quality,

informal costs, job satisfaction

For decades, criminologists have placed stock in employment as a policy

lever to deter criminal behavior, despite experimental evaluations failing to

consistently demonstrate joint improvements in employment and reductions

in offending at the individual level (Bushway and Apel 2012; but see Duwe

2012; Cook et al. 2015; Heller 2014; Visher, Debus, and Yahner 2008).

Nonexperimental research focusing on whether an individual was employed

within a given time period (measuring employment dic hotomously) has

similarly failed to show conclusive evidence that work deters crime.

1

This

lack of consistent support for an inverse work–crime relationship has led

scholars to question why work doesn’t work (Bushway and Apel 2012).

2

One explanation may be that the group of individuals that society is most

often interested in deterring—high-risk individuals with past criminal

involvement—do not frequently obtain the types of jobs which reduce

offending. Cook et al. (2015:378), for instance, questioned whether “a

temporary low-wage menial job” is a strong enough “treatment” to reduce

crime, suggesting that reducing recidivism may, in fact, “require a bigger

‘dose’ of legitimate opportunity.”

The sentiment that those with past criminal involvement would be set

straight if only they were able to acquire good jobs is also strongly backed

by criminological theory. Rational choice theory (RCT; Becker 1968; Cor-

nish and Clarke 1987) and theories of informal social control (Hirschi 1969;

Laub and Sampson 2003; Sampson and Laub 1993), for instance, propose

that a high-quality job will disincentivize criminal behavior. Given the

inverse relationship between job quality and crime posited, it is of no

surprise that researchers have called for deeper inquiry into the vast hetero-

geneity in job quality which has often been masked under the simple guise

of “employment” (Nguyen and Loughran 2018). More specifically, because

a good job and a bad job are not theoreticall y equivalent “treatments,”

Nguyen and Loughran (2018) highlighted that they should not be treated,

694 Journal of Research in Crime and Delinquency 57(6)

methodologically, as if they are. Rather, variation in job quality should be

conceptualized and measured as a key form of treatment heterogeneity

when evaluating the employment–crime relationship. However, when stud-

ies have moved beyond the “employment dichotomy” to consider treatment

heterogeneity, they have often implemented theoretically weak measures of

job quality or relied on only a single indicator. Due to these limitations,

previous research is frequently unable to provide insight as to whether

findings are substantively meaningful or merely the result of poor and/or

inconsistent operationalizations of quality.

A second explanation for inconsistent findings within the employment–

crime literature may be that improvements in job quality may not have the

same effect on offending for all types of people or across different places

and contexts. The notion that subgroups respond differently to an equivalent

treatment is referred to as treatment effect heterogeneity (Angrist 2004; see

also Nguyen and Loughran 2018). The life course tradition is well-known

for suggesting that age, place, and context can alter the role of turning points

in shaping one’s offending trajectory (e.g., Elder 1998; Sampson and Laub

1993). There is also a large criminological literature which suggests that

individuals may be differentially deterred (see Piquero et al. [2011] for an

overview). While research focused on the relationship between employ-

ment and crime has identified some instances of treatment effect hetero-

geneity, where the association between employment and crime differs by

key sociodemographic factors such as age, sex, race, and economic state

(Piquero, MacDonald, Parker 2002; Simons et al. 2002; Uggen 2000; Yang

2017), studies have often failed to systematically compare findings across

subgroups. Therefore, it is of little surprise that Nguyen and Loughran

(2018:349) also expressed that considering treatment effect heterogeneity

should be a “central concern” among criminologists.

This article aims to evaluate of the relationship between employment and

crime through a holistic evaluation of both treatm ent heterogeneity and

treatment effect heterogeneity. Using a sample of young adults with a

history of serious criminal involvement, the study examines treatment het-

erogeneity by implementing a theoretically grounded perceptual measure of

job quality (job satisfaction) and comparing findings across alternative

operationalizations of quality found in prior research (i.e., income, benefits,

hourly status, and job market sector). The study then assesses treatment

effect heterogeneity by considering whether the relationship between job

quality and offending is consistent across key sociodemographic character-

istics, individual differences, and social/situational contexts. The following

Jaynes 695