DO CELLMATES MATTER? A CAUSAL TEST OF THE SCHOOLS OF CRIME HYPOTHESIS WITH IMPLICATIONS FOR DIFFERENTIAL ASSOCIATION AND DETERRENCE THEORIES

AuthorHEATHER M. HARRIS,KIMINORI NAKAMURA,KRISTOFER BRET BUCKLEN
Published date01 February 2018
DOIhttp://doi.org/10.1111/1745-9125.12155
Date01 February 2018
DO CELLMATES MATTER? A CAUSAL TEST OF THE
SCHOOLS OF CRIME HYPOTHESIS WITH
IMPLICATIONS FOR DIFFERENTIAL ASSOCIATION
AND DETERRENCE THEORIES
HEATHER M. HARRIS,1KIMINORI NAKAMURA,2
and KRISTOFER BRET BUCKLEN3
1University of California, Berkeley
2University of Maryland—College Park
3Pennsylvania Department of Corrections
KEYWORDS: social interaction effects, prisonization, operant conditioning, local
instrumental variables, administrative data
In the schools of crime hypothesis, social interactions between inmates are assumed
to produce criminogenic rather than deterrent prison peer effects, thus implicating them
in the persistence of high recidivism rates and null or criminogenic prison effects. We
assess the validity of the schools of crime hypothesis by estimating prison peer effects
that result from differential cellmate associations in a male, first-time release cohort
from the Pennsylvania Department of Corrections. To isolate causal prison peer effects
in the presence of essential heterogeneity, we use a semiparametric local instrumental
variables estimation strategy. Our results do not support the school of crime hypoth-
esis. In our sample, prison peer effects produced in interaction with more criminally
experienced cellmates are always null or deterrent rather than criminogenic. Although
we do not explicitly test for the operant conditioning mechanisms theorized to under-
lie social influence in the context of differential association, we argue that, under the
assumption that the differential association context relates positively to the direction
of peer influence, our universally noncriminogenic estimates exclude direct reinforce-
ment, vicarious reinforcement, and direct punishment as potential drivers of prison
peer effects produced in interaction with more criminally experienced cellmates. Our
results support the assertion that operant conditioning mechanisms connect differential
association and deterrence theories.
Additional supporting information can be found in the listing for this article in the Wiley Online
Library at http://onlinelibrary.wiley.com/doi/10.1111/crim.2018.56.issue-1/issuetoc.
We thank the following individuals who contributed their expertise: Nicolette Bell at the Pennsyl-
vania Department of Corrections fulfilled our data requests and aided our understanding of the
data elements. Peter Reuter directed the dissertation on which this work is based. Laura Dugan
provided additional guidance. Dissertation committee members Sergio Urz´
ua, Terence Thorn-
berry, and Jean McGloin offered many helpful comments, as did David Harding, Aaron Cicourel,
Mauri Matsuda, and the participants at the Institute for Research on Poverty Summer Workshop,
particularly David Bjerk. The Graduate Research Fellowship Program at the National Institute of
Justice (2013-IJ-CX-0043) funded this work.
Direct correspondence to Heather M. Harris, Sociology Department, University of California,
Berkeley, Berkeley, CA 94720 (e-mail: hmh@berkeley.edu).
C2017 American Society of Criminology doi: 10.1111/1745-9125.12155
CRIMINOLOGY Volume 56 Number 1 87–122 2018 87
88 HARRIS, NAKAMURA, & BUCKLEN
Since the advent of criminological inquiry into prison effects, social interactions be-
tween inmates have been assumed to create capable criminals, rather than conscientious
citizens (Bentham, 1830). Put another way, prison peer effects are generally assumed to
be criminogenic rather than deterrent. To support this assumption, Clemmer (1950) ap-
plied Sutherland’s (1947) differential association theory to the prison environment, ar-
guing that by associating with more experienced criminals, prison inmates become in-
creasingly “prisonized” or socialized to antisocial prison norms, which leads to escalation
of their postrelease criminal activities. As a derivative of differential association, pris-
onization is expected to unfold in the context of differential association (Akers, 2009;
McGloin, 2009) and to operate through operant conditioning mechanisms (Burgess and
Akers, 1966; Stafford and Warr, 1993).
To modernize and expand on Clemmer’s (1950) framework, we consider how direct and
vicarious reinforcement and direct and vicarious punishment might produce criminogenic
and deterrent prison peer effects in a conventional differential association framework that
we connect to deterrence theory. By building on the work of Stafford and Warr (1993),
we argue that each of the four aforementioned operant conditioning mechanisms can the-
oretically produce criminogenic or deterrent prison peer effects. In the absence of direct
observation, additional assumptions are required to identify the underlying mechanisms.
Differential association theory hinges on the foundational assumption that the relative
differential association context relates positively to the direction in which peers exert in-
fluence (Akers, 2009; McGloin, 2009; Sutherland, 1947). That assumption limits which
mechanisms can be expected to produce which effects. Under it, reinforcement can pro-
duce criminogenic, but not deterrent, prison peer effects. Only by relaxing it can we ex-
pect more criminally experienced cellmates to reinforce lesser criminal behavior in their
prison peers.
To evaluate the schools of crime hypothesis, we sourced administrative data from the
Pennsylvania Department of Corrections (PADOC) and the Pennsylvania State Police
(PSP) to create a unique data set that characterizes the male members of a first-time
prison release cohort (N=10,116) and their longest cellmate associations. The method-
ological framework we develop enables us to isolate the causal effect of celling with
more criminally experienced cellmates on the reoffending of released prisoners. In the
Pennsylvania prison system, inmates can choose their cellmates. Self-selection into cell-
mate associations to affect reoffending creates the potential for essential heterogeneity to
bias our effect estimates. We account for essential heterogeneity with a local instrumental
variables estimation strategy (Heckman, Urz ´
ua, and Vytlacil, 2006).
We find no support for the schools of crime hypothesis. For each first-time inmate in
our sample who celled with a more criminally experienced longest cellmate, we find only
evidence of null or deterrent marginal prison peer effects. Similarly, our average prison
peer effect estimates are substantial in magnitude, significant, and deterrent. We discuss
our findings in relation to the framework we develop to unify differential association and
deterrence theories.
PRISON PEERS AND THEIR EFFECTS IN THEORY
AND PRACTICE
Bentham’s (1830: §VII) warning that “the indiscriminate association of prisoners”
can transform prisons from “places for reform” into “schools of crime” has led to the
DO CELLMATES MATTER? 89
expectation that prison peer effects are at least partly responsible for the failure of pris-
ons to reduce reoffending. For example, in their review of the incarceration and reoffend-
ing literature, Nagin, Cullen, and Jonson (2009: 124–64) cited the centuries-old schools of
crime hypothesis as one mechanism through which prisons might exert “null or crimino-
genic” effects on reoffending. Their application of the schools of crime hypothesis is not
unique. Since the midpoint of the past century, it has permeated the criminological ethos
with respect to the perpetuation and escalation of criminality among the formerly incar-
cerated (Lerman, 2009; Mears et al., 2013; Shaw, 1966).
A plausible theoretical rationale for the presence of criminogenic prison peer effects
invokes peer influence through social learning mechanisms, in other words, Sutherland’s
(1947) differential association theory, into which Burgess and Akers (1966) integrated op-
erant conditioning mechanisms and principles (Skinner, 1938, 1953). Within differential
association theory, the differential association context is assumed to predict the direction
of peer influence (Akers, 2009). Specifically, Sutherland (1947) argued that criminality,
or the underlying tendency to engage in criminal behavior, emerges and develops in in-
teraction with others who hold deeper criminal values and have greater criminal skills.
Therefore, in a differential association context characterized by the level of criminal ex-
perience one holds, those with more criminal experience promote criminal behavior in
those with lesser criminal experience.
In applying Sutherland’s theory to the prison context, Clemmer (1940, 1950) argued
that differential associations with other inmates lead to differential degrees of assimilation
to the prison milieu via prisonization, a normative socialization process that exacerbates
criminality by instilling antisocial norms. He expected prisonization to occur specifically
through social interactions with cellmates, a prediction supported by Gold and Osgood
(1992), who found that peer effects were most likely to arise between cellmates in juvenile
facilities in Michigan.
According to Clemmer (1950: 317), “a chance placement with a cellmate” with more
criminal experience instigates a cascade of events whereby an inmate is initiated into the
prison environment through his cellmate and his cellmate’s social contacts in the prison.
The well-established finding of peer similarity or homophily across multiple dimensions,
including criminality and criminal experience (Glueck and Glueck, 1950), suggests that a
cellmate with a lengthier criminal record is likelier to associate with other prisoners with
lengthier criminal records, thus, exposing the inmate to individuals whose criminality is
more likely to exceed and then augment his own via “relational mechanisms” (Kreager
et al., 2016: 8). From these early cellmate social ties, which are often short-lived as pris-
oners acclimate to the prison context in relationship fits and starts, cellmate associations
can develop and endure over long periods of time (Dishion, Veronneau, and Myers, 2010;
Earley, 2000; Jones and Schmid, 2000).
Prison peer effects resulting from differential cellmate associations might also be de-
terrent. McGloin (2009) argued that whether prohibited behavior increases or decreases
after social interactions in a context of differential association depends on the relative
distance between the criminality and the criminal experience of the interacting individ-
uals. As people seek balance in their relationships, the behavior of each moderates to-
ward that of the other (Heider, 1958). Thus, the differential association context becomes
arelative differential association context that continues to predict the direction of peer
influence. More criminal peers promote criminality, whereas less criminal peers deter it.
In her supporting analysis, McGloin (2009) found that paired peers became more similar

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT