A THRESHOLD MODEL OF COLLECTIVE CRIME

• Published date: 01 August 2015
• Author: ZACHARY R. ROWAN,JEAN MARIE MCGLOIN
• DOI: http://doi.org/10.1111/1745-9125.12077
• Date: 01 August 2015

A THRESHOLD MODEL OF COLLECTIVE CRIME∗

JEAN MARIE MCGLOIN and ZACHARY R. ROWAN

Department of Criminology and Criminal Justice, University of Maryland

KEYWORDS: group crime, threshold, co-offending

The group nature of offending has been recognized as an inherent characteristic of

criminal behavior, yet our insight on the decision to engage in group crime is limited.

This article argues that a threshold model offers broad appeal to understand this deci-

sion. After discussing the basis of this model and its applicability to collective crime, we

offer one example of the kind of research that could stem from this model. Specifically,

by using survey data from 583 university students, this study asked respondents to self-

report thresholds for group theft and destruction of property. By experimentally ma-

nipulating characteristics of the hypothetical scenario used to measure thresholds, we

investigated both the individual- and situational-level correlates of these self-reported

thresholds. The discussion considers the results that emerge from a Tobit regression

model and offers suggestions for future research that would provide further refinement

of the threshold model.

Notable growth in empirical attention to group crime has occurred in recent years; re-

searchers have contributed to knowledge on the differential tendencies to offend with

others (e.g., Carrington, 2009; McGloin et al., 2007), the structure of co-offending groups

(e.g., McGloin and Piquero, 2010; Sarnecki, 2001), and the factors that distinguish those

who instigate group crime from those who follow others into the act (e.g., McGloin and

Nguyen, 2012; van Mastrigt and Farrington, 2010). Yet, perhaps the most fundamental

issue—the process whereby people decide to engage in group crime and the factors that

influence this decision—remains underdeveloped (McCarthy, Hagan, and Cohen, 1998;

Nguyen and McGloin, 2013). Given that delinquency often is group behavior (Erickson

and Jensen, 1977; Reiss, 1988), one can argue that if the discipline seeks a fuller under-

standing of the decision to engage in crime, then it also must consider the decision to

engage in group crime (Hochstetler, 2001).

Simply applying traditional rational choice models to this decision process does not

seem to be the solution. As McCarthy, Hagan, and Cohen (1998: 158) observed, such

models do not always incorporate the notion that “the probabilities associated with var-

ious outcomes are not independent; indeed, they are often highly interdependent” (see

also McCarthy, 2002). In other words, choices about whether to engage in group crime

may depend, at least in part, on the observed or anticipated actions of others (see also

Alarid, Burton, and Hochstetler, 2009; Hochstetler, 2001). The two extant theories about

the decision to co-offend embrace this idea of a conditional decision process (McCarthy,

∗The authors wish to thank Dr. Christopher Sullivan, Dr. Laura Dugan, Dr. Kyle Thomas, Dr. Holly

Nguyen, and Dr. Denise Gottfredson for their help and advice on earlier versions of this article. Di-

rect correspondence to Jean Marie McGloin, 2220L LeFrak Hall, University of Maryland, College

Park, MD 20707 (email: jmcgloin@umd.edu).

C2015 American Society of Criminology doi: 10.1111/1745-9125.12077

CRIMINOLOGY Volume 53 Number 3 484–512 2015 484

A THRESHOLD MODEL OF COLLECTIVE CRIME 485

Hagan, and Cohen, 1998; Weerman, 2003), but they are limited in at least one impor-

tant dimension. Specifically, these theories seem to be interested in explaining why peo-

ple who are already inclined toward offending take on accomplices (see also Tremblay,

1993). This focus, perhaps unintentionally, ignores one of the most compelling features of

group crime: “[P]eople will commit acts when they are with others that they would never

have committed if they had been alone” (Warr 2002: 59, emphasis in original; see also

Diener et al., 1976; Festinger et al., 1952). Ideally, a theory about the decision to engage

in group crime should be able to explain why individuals who normally may be motivated

for crime choose to co-offend as well as why people who normally would not engage in

such behavior decide to take part.

The key to such an explanation may rest in embracing the idea that group crime is one

form of collective behavior; this label includes a wide range of actions, such as the adop-

tion of innovation, migration trends, and voting patterns (Matsueda, 2006; Warr, 2002).

Thinking about group crime under this framework builds our understanding by “stealing

from our friends” in other disciplines that have long wrestled with understanding and ex-

plaining collective action (Osgood, 1998). Interestingly, one seminal model of collective

action draws attention to a fundamental characteristic of group crime that has typically

been portrayed as merely descriptive, as opposed to instrumental: the presence of others.

Granovetter’s (1978) threshold model asserted that an individual’s decision of whether

to engage in collective behavior is conditional on the actions of those around him or her.

Specifically, he proposed that individuals have varying “thresholds,” which describes the

proportion of people who must partake in the behavior before they join in. These thresh-

olds essentially act as tipping points in decision making, serving as the point at which

an individual comes to believe that partaking in the action will now maximize his or her

utility.

Despite the prominence of threshold models in the sociological literature, criminolog-

ical research leveraging this concept has been sparse (e.g., McGloin and Piquero, 2009).

Many behaviors to which threshold models have been applied are admittedly different

from offending, but Granovetter was clear that his model should apply to instances of

group crime—indeed, he relied on riot behavior to illustrate his points. In light of the

potential gains that can be earned by understanding the decision to partake in collective

crime, this article considers how a threshold model applies to group crime. In doing so,

it argues that this model can capture a wide range of co-offending events, accommodate

facts about co-offending, and afford the opportunity to consider, in a systematic way, how

individual-level and situational-level factors shape the decision to engage in group crime.

Then, the article offers an example of the kind of research that this theory may prompt,

broadening the traditional empirical approach to co-offending studies.

UNDERSTANDING THE DECISION TO ENGAGE IN

COLLECTIVE CRIME

Scholars have long noted that people behave differently when in a group setting as

opposed to when alone, often demonstrating a greater proclivity for “riskier” decisions

in the former context (e.g., Gardner and Steinberg, 2005). Early theorizing on collective

behavior suggested this stemmed from people losing their capacity for individual reason,

finding themselves transformed through some type of collective consciousness (Freud,