ON THE RELIABILITY AND VALIDITY OF SELF‐REPORTED ILLEGAL EARNINGS: IMPLICATIONS FOR THE STUDY OF CRIMINAL ACHIEVEMENT*

• DOI: http://doi.org/10.1111/1745-9125.12144
• Published date: 01 August 2017
• Author: THOMAS A. LOUGHRAN,HOLLY NGUYEN
• Date: 01 August 2017

ON THE RELIABILITY AND VALIDITY

OF SELF-REPORTED ILLEGAL EARNINGS:

IMPLICATIONS FOR THE STUDY OF CRIMINAL

ACHIEVEMENT∗

HOLLY NGUYEN1and THOMAS A. LOUGHRAN2

1Pennsylvania State University

2University of Maryland—College Park

KEYWORDS: illegal earnings, criminal achievement, reliability, validity

The study of the monetary returns to criminal activity is a central component in

many emerging areas of criminology, including rational choice and offender decision-

making, desistance, and criminal achievement. Scholars have been increasingly capti-

vated with specification of the earnings function and with examining how variations

in illegal earnings predict important outcomes such as persistence in offending. The

potential utility of findings in related empirical studies hinges on the quality of the key

measure, self-reported illegal earnings. Yet to date scant attention has been paid by re-

searchers to the measurement properties of this metric. We analyze self-reported illegal

earnings generated from a variety of instrumental crimes by using data from the Path-

ways to Desistance Study (n =585) and the National Supported Work Project (n =

1,509), which are two longitudinal data sets of active offenders separated by more than

30 years. Findings based on analyses both within and between data sets reveal support

for the internal consistency reliability and criterion validity of self-reported illegal earn-

ings. Moreover, the results reveal premiums in terms of higher earnings associated with

different crime types, which are persistent both over time and across data sets. Implica-

tions and future directions for advancing the theoretical study of criminal achievement

are also discussed.

Instrumental, or economically motivated, crimes are directly harmful to victims, are

seven times more prevalent than violent crimes, and involve billions of dollars of eco-

nomic losses annually (Baumer and Gustafson, 2007; Messner and Rosenfeld, 2001;

UNODC, 2013). Baumer and Gustafson (2007: 632) defined these types of crimes as of-

fending “geared primarily towards the acquisition of money or goods that could be con-

verted to cash.” As such, a motivating factor for the participation in, frequency of, and

type of offending is the illegal earnings gained from such activities.

Studying illegal earnings is important to many fundamental issues in sociology, crim-

inology, and economics. Illegal earnings are one of the most tangible rewards to crime

and play an important role in offender decision-making. As such, several scholars have

∗We would like to thank Wayne Osgood, Terry Thornberry, Darrell Steffensmeier, Brandy Parker,

Martin Bouchard, Rosemary Gartner, and the anonymous reviewers for their thoughtful comments

on earlier drafts of this article.

Direct correspondence to Holly Nguyen, Department of Sociology and Criminology, The

Pennsylvania State University, 905 Oswald Tower, University Park, PA 16802 (e-mail:

hollynguyen@psu.edu).

C2017 American Society of Criminology doi: 10.1111/1745-9125.12144

CRIMINOLOGY Volume 55 Number 3 575–602 2017 575

576 NGUYEN & LOUGHRAN

attended to the specification of the illegal earnings function (e.g., Loughran et al., 2013;

Nguyen et al., 2016; Uggen and Thompson, 2003). The study of illegal earnings also pro-

vides insight into criminal achievement and competence, which have implications for not

only offender decision-making and rational choice theories of crime but also persistence

in and desistance from criminal careers. For instance, McCarthy and Hagan (2001: 1036)

noted that “offenders vary considerably in their illegal accomplishments” and that factors

such as criminal social capital are influential to success in illegal endeavors and are im-

portant criminal analogs to factors influential to conventional forms of success. The illicit

rewards from crime have also been linked to involvement in economic crimes (Pezzin,

1995; Sampson and Laub, 2003; Steffensmeier, 1986), desistance processes (Uggen and

Kruttschnitt, 1998), perceptions of criminal self-efficacy (Laferriere and Morselli, 2015),

and criminal achievement, which may be ultimately linked to persistence in criminal

involvement (McCarthy and Hagan, 2001; Morselli, Tremblay, and McCarthy, 2006;

Shover and Thompson, 1992). Additionally, the results of studies aimed at examining

illegal earnings have revealed associations with involvement in the drug trade and the

effectiveness of drug treatment (French and Zarkin, 1992; Reuter et al., 1990; Yule, Par´

e,

and Gartner, 2014).

Despite the depth of interest in this key measure, almost all studies aimed at examin-

ing illegal earnings have relied on self-reported data. As such, it is both reasonable and

necessary to question the accuracy of self-reported illegal earnings measures. To date,

however, there has been little inquiry into the reliability and validity of these measures.

As scholars continue to deploy increasingly sophisticated statistical models to study ille-

gal earnings, concerns about measurement intensify. For instance, Angrist and Krueger

(1999) warned that classic measurement error is generally exacerbated in panel data mod-

els such as difference estimators. Several studies have been focused on examining within-

individual changes in self-reported illegal earnings over time (i.e., Nguyen et al., 2016;

Thompson and Uggen, 2012; Uggen and Thompson, 2003).

In this study, we move beyond prior research by providing detailed analyses of the

measurement of self-reported illegal earnings. Specifically, we use two different, although

complementary, sources of longitudinal data collected more than 30 years apart—the

Pathways to Desistance Study and the National Supported Work (NSW) Demonstration

Project—to study the internal consistency reliability and criterion validity of self-reported

illegal earnings. The results of our analyses yield both methodological and substantive

findings. With regard to the former, our results demonstrate correspondence between

different measures of self-reported illegal earnings both within data sets and across data

sets. For the latter, we find evidence of higher (real) illegal premiums for participation in

similar crime types across time and show that returns to participation in different crimes

are consistently rank ordered.1Finally, we consider implications for future study of illegal

earnings and criminal achievement.

ASSESSING SELF-REPORTED ANTISOCIAL BEHAVIOR

To begin to study the key measurement properties of illegal earnings, we draw from

two parallel yet distinct literatures that also typically comprise self-reports—research on

1. To be clear, we use the term premium throughout this article to refer to the additional, higher

return for participation in one activity as compared with another.