Replication in criminology: A necessary practice
| Published date | 01 September 2015 |
| Date | 01 September 2015 |
| Author | Susan McNeeley,Jessica J. Warner |
| DOI | 10.1177/1477370815578197 |
| Subject Matter | Articles |
European Journal of Criminology
2015, Vol. 12(5) 581 –597
© The Author(s) 2015
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1477370815578197
euc.sagepub.com
Replication in criminology:
A necessary practice
Susan McNeeley
Pennsylvania State University, USA
Jessica J. Warner
University of Cincinnati, USA
Abstract
Although researchers acknowledge the importance of replication in building scientific knowledge,
replication studies seem to be published infrequently. The present study examines the extent
to which replications are conducted in criminology. We conduct a content analysis of the five
most influential journals in criminology. We also compare the replication rate in criminology
with that in the social sciences and natural sciences. The results show that replication research
is rarely published in these disciplines. In criminology journals in particular, replication studies
constitute just over 2 percent of the articles published between 2006 and 2010. Further, those
replication studies that were published in criminology journals in that period tended to conflict
with the original studies. These findings call into question the utility of empirical results published
in criminology journals for developing theory and policy. Strategies for promoting replication
research in criminology are suggested.
Keywords
Content analysis, quantitative methods, replication, research methods
Introduction
Replication is an essential component of the scientific process. Scholars agree that the
results of any empirical study cannot be fully trusted until they have been replicated in
additional studies. The concept of replication is so familiar to the scientific community that
its meaning seems obvious, making formal, agreed-upon definitions rare. For example,
Corresponding author:
Susan McNeeley, Justice Center for Research, Pennsylvania State University, 327 Pond Building, University
Park, PA 16802, USA.
Email: smm70@psu.edu
578197EUC0010.1177/1477370815578197European Journal of CriminologyMcNeeley and Warner
research-article2015
Article
582 European Journal of Criminology 12(5)
Schweizer (1989) compiled a list of definitions for replication in the social sciences. Most
of these include a component regarding the reproduction of an experimental procedure;
additionally, many definitions stress the use of different, independently collected data.
It is important to note that not all studies should necessarily be expected to replicate.
For example, ethnographies and other qualitative work or work with a historical focus
often examine specific groups and contexts without making claims about the generaliz-
ability of any observations made. Certainly, some criminological research meets these
criteria and should not be expected to be subjected to replication. However, a great deal
of criminology research is meant to test general theories or evaluate interventions and
initiatives, and these types of studies are in need of replication. Additionally, because
criminology is an applied field, findings from research studies may be applied by crimi-
nal justice agencies. Policy and practice should not be guided by information that has not
been properly validated. Therefore, the greater degree of confidence that is provided by
replications is needed. For this reason, replication of highly publicized studies is even
more important, as such research is more likely to come to the attention of policymakers
and practitioners. Further, replication would allow practitioners to have more confidence
in the original findings, lending credibility to the research process and making the inte-
gration of research results into policies and practice more attractive.
There have been important replications conducted in criminology research that shed
light on the importance of replicating findings before making policy implications. First,
early studies showed that arrests were more likely to occur when police arrived at a crime
scene shortly after a call for service is made (Clawson and Chang, 1977; Isaacs, 1967).
However, subsequent replications (for example, Pate et al., 1976; Spelman and Brown,
1981) failed to find a significant relationship between response times and arrest, or found
that the relationship held for only certain types of crime, such as those involving direct
contact between the victim and offender. Second, in the Minnesota Domestic Violence
study, Sherman and Berk (1984) examined the effectiveness of multiple police responses
to domestic violence calls. Their findings showed that offenders who were arrested were
less likely to reoffend than were those who were given counseling or temporarily sent
away. Several replications of the study, some of them NIJ-funded, have since been con-
ducted. Many of these failed to support the findings of the original experiment (for
example, Dunford, 1992; Hirschel et al., 1992; Sherman et al., 1991); those that did find
results in the same direction reported weaker relationships between arrest and repeat
offending than did the original study (for example, Berk et al., 1992; Pate et al., 1991).
These results demonstrate the potential unreliability of single studies; the findings of
some studies may be sensitive to the context in which they are conducted or the specific
research methods used. This highlights the importance of replicating criminological
studies in an effort to provide more reliable results on which to base policy and practice.
Although these examples are well known and the importance of replication is largely
agreed upon, it is currently unknown whether the results reported in criminology journals
suffer from a lack of replication.
Owing to the importance of replication and the concern over the frequency of replica-
tion research, we conduct an exploratory study that examines the number of replication
studies that have recently been published in top-tier criminology and criminal justice
journals. By doing so, we hope to provide initial insight regarding the extent to which the
To continue reading
Request your trial