Application of bifactor models in criminal psychology research: a guide to researchers

Publication Date05 May 2015
AuthorPhilip Hyland
SubjectHealth & social care,Criminology & forensic psychology,Criminal psychology
Application of bifactor models in criminal
psychology research: a guide
to researchers
Philip Hyland
Dr Philip Hyland is Lecturer in
Psychology at the School of
Business, National College of
Ireland, Dublin, Ireland.
Purpose The purpose of this paper is to introduce the reader to the nature of confirmatory bifactor modelling.
Confirmatory bifactor modelling is a factor analytic procedure that allows researchers to model unidimensionality
and multidimensionality simultaneously. This method hasimportant appli cationsin the field of criminal psychology.
Design/methodology/approach This paper begins by introducing the topic of factor analysis and
explains how confirmatory bifactor modelling is similar yet distinct to the more familiar factor analytical
procedures in the psychological literature.
Findings Through practical examples this paper explains the value of this analytical technique to
researchers in criminal psychology. Examples from the existing criminal psychological literature are used
to illustrate the way in which bifactor analysis allows important theoretical questions to be addressed.
Originality/value This paper highlights the strengths and limitations associated with traditional restricted
confirmatory bifactor models and introduces the notion of the unrestrictedbifactor model. The unrestricted
bifactor model allows greater flexibility for addressing interesting research questions. The paper concludes by
providing the reader with an annotated Mplus syntax file for how to perform confirmatory bifactor modelling.
Keywords Psychopathy, Bifactor, Confirmatory factor analysis (CFA), Criminal social identity,
Latent variable modelling, Structural equation modelling (SEM)
Paper type Conceptual paper
Factor analysis
The use of latent variable modelling techniques in the social sciences ha s greatly improved the
scientific integrity of these disciplines. Application of latent variable modelling techniques such as
factor analysis and structural equation modelling (SEM) have allowed social science researchers to
more precisely measure their target variables of interest (through removal of measurement error)
and to develop more parsimonious theoretical models (through the construction of latent variables).
Factor analytical procedures, whether exploratory or confirmatory, have substantial value to
the scientifically minded criminal psychologist. Criminal psychologists who seek to understand the
factors that contribute to the development and maintenance of criminal behaviour often investigate
numerous psychological constructs to explain these actions: psychopathy (Hare, 1991); personality
traits (Eysenck and Gudjonsson, 1989); self-esteem (Boduszek et al., 2012a, b); criminal thinking
styles (Walters, 2012); and criminal social identity (Boduszek and Hyland, 2011). All of these
constructs, and every other psychological variable, is termed a latent variablegiven that it cannot
be directly observed. Latent variables are distinguished from observed variableswhich, as the
name suggests, are directly observable (e.g. type of criminal conviction, amount of time spent in
prison, number of criminal convictions).
Received 10 March 2015
Revised 16 March 2015
Accepted 18 April 2015
DOI 10.1108/JCP-03-2015-0011 VOL. 5 NO. 2 2015, pp. 65-74, © Emerald Group Publishing Limited, ISSN 2009-3829
PAG E 65

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