Multidimensional Latent Classification of ‘Street Robbery’ Offences

In a recent study of personal robbery, commissioned by the Home Office in the UK, a qualitative typology of robbery offences was proposed based on the approach used by the offender to commit the crime, consisting of four approach types: Blitz, Confrontation, Con, and Snatch. Conceptual inspection of the typology reveals that these proposed types may be hypothetically demarcated as the product of two latent dimensions: interaction (between the offender and the victim) and violence (used to threaten/harm the victim). The current paper utilises crime scene information from 72 incarcerated male offenders convicted of ‘street’ robbery to test this hypothesis. Convergent statistical analysis was utilised to test the structure of Smith's typology first using multidimensional scaling (MDS) and then principal component analysis (PCA). MDS and PCA analyses provided convergent support for the existence of the four robbery styles and the latent dimensions of interaction and violence. Implications of Smith's typological structure and latent behavioural dimensions on the conceptualisation and classification of robbery offences are discussed within the existing literature on ‘street’ robbery. Copyright © 2012 John Wiley & Sons, Ltd.     

Comparing the Predictive Accuracy of Case Linkage Methods in Serious Sexual Assaults

The empirical support for linkage analysis is steadily increasing, but the question remains as to what method of linking is the most effective. We compared a more theory‐based, dimensional behavioural approach with a rather pragmatic, multivariate behavioural approach with regard to their accuracy in linking serial sexual assaults in a UK sample of serial sexual assaults (n = 90) and one‐off sexual assaults (n = 129). Their respective linkage accuracy was assessed by (1) using seven dimensions derived by non‐parametric Mokken scale analysis (MSA) as predictors in discriminant function analysis (DFA) and (2) 46 crime scene characteristics simultaneously in a naive Bayesian classifier (NBC). The dimensional scales predicted 28.9% of the series correctly, whereas the NBC correctly identified 34.5% of the series. However, a subsequent inclusion of non‐serial offences in the target group decreased the amount of correct links in the dimensional approach (MSA–DFA: 8.9%; NBC: 32.2%). Receiver operating characteristic analysis was used as a more objective comparison of the two methods under both conditions, confirming that each achieved good accuracies (AUCs = .74–.89), but the NBC performed significantly better than the dimensional approach. The consequences for the practical implementation in behavioural case linkage are discussed. Copyright © 2012 John Wiley & Sons, Ltd.     

Two-Way Regularized Fuzzy Clustering of Multiple Correspondence Analysis

Multiple correspondence analysis (MCA) is a useful tool for investigating the interrelationships among dummy-coded categorical variables. MCA has been combined with clustering methods to examine whether there exist heterogeneous subclusters of a population, which exhibit cluster-level heterogeneity. These combined approaches aim to classify either observations only (one-way clustering of MCA) or both observations and variable categories (two-way clustering of MCA). The latter approach is favored because its solutions are easier to interpret by providing explicitly which subgroup of observations is associated with which subset of variable categories. Nonetheless, the two-way approach has been built on hard classification that assumes observations and/or variable categories to belong to only one cluster. To relax this assumption, we propose two-way fuzzy clustering of MCA. Specifically, we combine MCA with fuzzy k-means simultaneously to classify a subgroup of observations and a subset of variable categories into a common cluster, while allowing both observations and variable categories to belong partially to multiple clusters. Importantly, we adopt regularized fuzzy k-means, thereby enabling us to decide the degree of fuzziness in cluster memberships automatically. We evaluate the performance of the proposed approach through the analysis of simulated and real data, in comparison with existing two-way clustering approaches.     

Behavioural linking of stranger rapes

A sample of serial stranger rape cases ( n = 43) that had occurred in Finland during the years 1983–2001 were studied with the objectives being to: (a) describe the characteristics of the offenders; (b) explore the structure of serial rape; and (c) demonstrate behavioural linkage through an analysis of the offenders' crime scene behaviour using both multidimensional scaling (MDS) and discriminant function analysis (DFA). The material was content analysed with regard to the occurrence of a number of dichotomous variables. The inter‐relationships of the variables was studied using MDS. The analysis revealed two previously identified major modes of interaction with the victim: involvement and hostility. Employing MDS and DFA, it was shown that the offences of different offenders were distinguishable in terms of variation between the offences of different offenders and consistency within the offences of a single offender. Using DFA, the classification accuracy clearly exceeds that expected by chance, and 25.6% of the cases were classified without any error. The results are discussed in relation to their practical utility and previous studies. Copyright © 2005 John Wiley & Sons, Ltd.     

Profiling Korean sex crimes: Offender characteristics and crime scene actions

The present study aimed to examine the demographic information of sex offenders in South Korea and explore whether a U.K. thematic model of criminal behaviour could be replicated in the Korean context. The 27 variables of crime scene actions derived from 50 Korean sexual offences through a content analysis were analysed with smallest space analysis. Chi‐square was also administered to explore the differences in offender characteristics among behavioural themes. Consequently, three separate action themes, hostility–involvement, theft, and control, were revealed. Next, each case was assigned to one of the dominant or hybrid themes, with 84% of the total cases being classified to the dominant themes. Moreover, there were significant differences in four offender characteristics between the themes: “previous conviction of robbery,” “previous conviction of sexual crime,” “knowing the victim,” and “vehicle use.” These outcomes have implications for the development of the Korean profiling system.     

Advancement of Criminal Profiling Methods in Faceted Multidimensional Analysis

The current study seeks to advance the faceted multidimensional scaling (termed FMDS) procedure used in much of the investigative psychology research. To this end, recent research on street robbery by Goodwill and colleagues will be utilised to illustrate the effectiveness of a facet scale method for offender profiling. Four FMDS themes of street robbery (Con, Blitz, Confrontation and Snatch) were revealed by the crossing of two underlying axial facets: the offenders' level of violence and interaction with the victim. The facet scale method, utilising offenders' axial facet scores, was compared to previous count, proportional and centroid classification methods in the prediction of offender criminal histories. Utilising logistic regression and receiver operating characteristic analyses, the axial facet scale method was found to significantly outperform the qualitatively based dominant theme classification methods that typically employ angular and radial facets for FMDS interpretation. Implications for the use of axial facet scales within FMDS analysis for offender profiling research are discussed. Copyright © 2012 John Wiley & Sons, Ltd.     

Serial homicide: an investigation of behavioural consistency

To examine if serial homicide offenders are consistent across their crimes, a model was developed empirically that could be used to distinguish between crimes that were instrumental and those that were expressive. The first known three offences in each series of 69 US serial homicides committed by 23 offenders, were examined and the instrumental and expressive themes determined. Three models were then explored that test for consistency across these themes. The most liberal model was found to classify all of the offences effectively and to reveal complete consistency across the three crimes for all offenders. The implications of these results for offender profiling and further study of serial homicide are discussed. Copyright © 2005 John Wiley & Sons, Ltd.     

Behavioural Profiles and Offender Characteristics across 111 Korean Sexual Assaults

Although the potential usefulness of the offence action–offender characteristic (A–C) relationships is widely accepted and operational ‘offender profiling’ units now exist around the world, few such relationships have been empirically established. To explore this, the offending action patterns within 111 sexual assault cases from South Korea were coded in terms of 16 distinctive, objective crime scene criteria and subjected to an agglomerative hierarchical cluster analysis. Background psychiatric and general characteristics, Personality Assessment Inventory (PAI) scale scores, and criminal histories were described for each cluster. The cluster analysis drew attention to six clusters or behavioural profiles within the sexual assaults. Cluster 1 included serial offenders who aggressively raped and robbed adult women, with some pseudo‐intimate sexual behaviour, in their homes. Two thirds of these offenders had histories of sexual assault. Cluster 2 included offenders who again targeted adults in their homes, but without pseudo‐intimate sexual behaviour. Cluster 3 included offenders who targeted adults outdoors at night. These offenders showed high antisocial personality PAI scores and psychiatric histories of sexual sadism. Cluster 4 included unarmed offenders who targeted adults in their homes without robbery. These offenders often had psychiatric histories of depression. Cluster 5 included offenders who targeted adults outdoors with a blitz‐style attack, and Cluster 6 included offenders who targeted minors outdoors, without weapons, using a confidence‐trick style of approach. Paedophilia and histories of psychiatric treatment were prominent amongst these offenders. The results indicate therefore some of the key empirical relationships that future research may develop as the basis for sexual assault ‘profiles’. Copyright © 2014 John Wiley & Sons, Ltd.     

UK and Spanish stranger sexual offenders crime scene behaviours and previous convictions: A cross-cultural comparison

International comparisons of previous convictions (PC) and crime scene behaviours (CSB) of stranger sexual offences can inform offender profiling strategies; especially the degree to which pragmatic models cross validate across countries. The present study compared PC and CSB of 474 UK and 418 Spanish cases. CSB and PC were analysed using Chi-square. UK stranger sexual offenders displayed a higher proportion of all CSB analysed. UK offenders also had a significantly higher proportion of PC with the exception of violence offences. There were no significant differences in sexual pre-convictions. In both the UK and Spain the CSB ‘reference to the police’ was significantly association with the PC ‘Criminal damage’ no other similar significant associations were found. Explanations for these findings include differences between the countries in regards to: legislation, crime recording processes, data coding processes, with additional consideration of any cultural, environmental and contextual factors.     

The role of offender profiling in classifying rapists: Implications for counselling

The offender profiling process is an investigative technique used to identify the major personality and behavioural characteristics of the offender based upon an analysis of the crime(s) he or she has committed (Douglas & Burgess, 1986). Attempts have been made in both England and America to classify rapists into various subgroups. Instead of definitive groups of offenders a suggestion is made as to the formation of a dimensional typology of rapists. This would allow for the difficulties in defining boundaries for subgroups and accounting for individual differences and heterogeneity of both the rapist and the assault. The various criteria that attempt to distinguish the different types of offender are discussed including behavioural and motivational characteristics of the offender, lay theories of rape and the victim-offender relationship. The implications of the dimensional typology are outlined. There are indications that some factors may contribute to a longer and more complicated recovery process. Macnab (1993) neatly outlines eleven contributory factors: financial stress, various long term stresses, the degree of violence, the use of weapons, the sense of violation, the threat of death, multiple assailants, threats of greater or continuing harm and a later death, medical complications, problems with police, medical and criminal systems and lack of adequate professional and informed support. As well as this, the important issue of self-blame is discussed along with the subsequent problem of post-traumatic stress disorder: the recurring distressing memories, dreams and events and signals that can provoke a further cycle of distress.     

Short-term predictive validity of cluster analytic and dimensional classification of child behavioral adjustment in school

A constructive debate over the classification of child psychopathology can be stimulated by investigating the validity of different classification approaches. We examined and compared the short-term predictive validity of cluster analytic and dimensional classifications of child behavioral adjustment in school using the Behavior Assessment System for Children (BASC; [Reynolds, C. R., and Kamphaus, R. W. (1992). Behavior Assessment System for Children (BASC). Circle Pines, MN: AGS.]) Teacher Rating Scales for Children (TRS-C). We cluster analyzed 14 dimensional scores of children's behavior in first grade and identified seven clusters. Then we examined the predictive power of the dimensions and the clusters using a variety of school outcomes in second grade, including academic achievement scores in reading and math, absenteeism, discipline reports, and participation rates in pre-referral intervention. Both methods significantly predicted all school outcomes with similar magnitudes of effect sizes, but the dimensional approach was more powerful in predicting the outcomes except pre-referral intervention. Moreover, coherent group differences in the clusters were identified with respect to the school outcomes. This latter finding may suggest differentiated prognoses of cluster membership. The implications for future validation studies of alternative classification systems of child behavioral adjustment are discussed.     

A Factor Mixture Model Approach to Elaborating on Offender Mental Health Classification With the MMPI–2–RF

A large proportion of prison inmates suffer from mental illnesses or severe personality disorders; therefore, offender classification is a worthwhile endeavor both for efficiently allocating mental health treatment resources and security risk classification. This study sought to elaborate on offender classification by using an advanced statistical technique, factor mixture modeling, which capitalizes on the strengths of both latent trait analysis and latent class analysis. A sample consisting of 616 male and 194 female prison inmates was used for this purpose. The MMPI–2–RF Restructured Clinical (RC) scales were used to elaborate on a variety of latent trait, latent class, and factor mixture models. A 3-factor, 5-class mixture model was deemed optimal in this sample. Remaining MMPI–2–RF scales as well as scores on external criterion measures relevant to externalizing psychopathology were used to further elaborate on the utility of the resulting latent classes. These analyses indicated that 3 of the 5 classes were predominantly different expressions of externalizing personality proclivities, whereas the remaining 2 indicated inmates with substantial internalizing or thought-disordered characteristics. Implications of these findings are discussed.     

South African Serial Homicide: A Victim‐Focused Behavioural Typology

This current study assessed the value of the Interpersonal Model in differentiating serial homicide offences in South Africa, notably in terms of how the offender may display behavioural patterns that are indicative of them dealing with the victim as a Person, an Object, or a Vehicle. The sample consisted of 302 offences committed by 33 offenders that occurred from 1953 to 2007 in South Africa. Multidimensional scaling analysis was used to test the Interpersonal Model and results indicated that crime themes did not directly correspond to the threefold model. Instead, two crime scene types were present: Victim as Object (where the focus was on tangible gains, interaction with the victim throughout the phases of the crime was limited, and victims included men and live women) and Victim as Vehicle (where the victim might be construed as a conduit through whom the offender could realise their specific psychological needs, the interaction was extensive, and victims tended to be vulnerable). South African serial homicides did not appear to have a specific sub‐theme of Victim as Person, suggesting that the themes engaged in centred more on instrumentally focused actions but in qualitatively different ways. However, victim types were integral to the overall behavioural model. Using the Victim as Object/Vehicle dichotomy, 85.7% of the offences could be seen to engage dominantly in one of the two patterns at the crime scene. This not only supported the results of a twofold model, but having these many offenders showing a dominant crime scene type also shows that the model is an excellent representation of serial homicide offending in South Africa. Copyright © 2014 John Wiley & Sons, Ltd.     

Statistical modelling in the investigation of stranger rape

A sample of stranger rape offences (n = 271) registered in the Dutch Violence Crime Linkage Analysis System database in the Netherlands between 1997 and 2007 was studied with the objective of developing statistical models, which give an indication of the probability of basic offender characteristics. Observable crime characteristics concerning the modus operandi, interaction between the offender and the victim, violence, precautionary measures, and sexual behaviours were selected in the dataset. Offender characteristics were selected based on their usefulness for the police organisation in narrowing the scope of a criminal investigation. Spatial behaviour, criminal history, and living situation of the offender were selected. From the predictive models, four out of five achieved a correct rate of over 70%, and all models predicted better than the best guess method. The proposed models for distance and prior convictions for violence seem particularly promising. Both these models show an improvement of correctly predicted offender characteristics of more than 20 percentile points compared with that which could have been estimated based on the average in the total sample. The predictive value of the models needs to be tested further with ‘new offences’, which were not used to construct the model. In general, the current study supports the finding that crime characteristics can be used to get an indication of the probability of certain offender characteristics. Nevertheless, for an understanding of the relationship between the crime characteristics and offender characteristics, a further development of a theoretical framework is urgently necessary. Copyright © 2009 John Wiley & Sons, Ltd.     

Victim Gender,Number of Perpetrators,and Interpersonal Interaction in Stranger Rape: An Analysis of Direct and Moderator Effects

This study examined the direct and moderator effects of number of perpetrators and gender of victim on interpersonal behaviour in stranger rape. Crime scene behaviours representative of hostility, involvement, control, and offender penetration in rape were examined for 496 UK, police‐recorded cases of stranger rape. Cases were grouped according to victim gender (male or female) and number of perpetrators (lone or multiple). This resulted in four groups (lone female, lone male, multiple female, and multiple male) with 124 cases in each. Binary logistic regression and one‐way analysis of variance were used to investigate the relationships between the two predictor variables and 11 criterion variables. Significant direct effects of number of perpetrators were found whereby multiple perpetrator offences were more likely to involve violence and less likely to involve involvement interactions than lone perpetrator offences. Significant direct effects of victim gender were also found whereby male victims were more likely than female victims to experience hostile interactions and be threatened with a weapon and were less likely to experience offender penetration and involvement interactions. Significant crossover interactions were also found for four hostility variables. The utility of the findings are discussed in relation to crime prevention, victim support, and offender intervention. Copyright © 2013 John Wiley & Sons, Ltd.     

Characterising the Personality of the Public Safety Offender and Non‐offender using Decision Trees: The Case of Colombia

The aim of this paper was to create a decision tree (DT) to identify personality profiles of offenders against public safety. A technique meeting this requirement was proposed that uses the C4.5 algorithm to derive decision rules for personality profiling of public safety offenders. The Mini‐Mult test was used to measure the personality profiles of 238 individuals. With the test results as our database, a C4.5 DT was applied to construct rules that classify each profile into one of two groups, those without and those with records of offences against public safety. The model correctly classified 80% of the personality profiles and delivered a set of decision rules for distinguishing the profiles by group, and the principal personality profiles were interpreted. We conclude that DTs are a promising technique for analysing personality profiles by their offender or non‐offender status. Finally, we believe that the development of a classifying model using DT may have practical applications in the Colombian prison system. Copyright © 2016 John Wiley & Sons, Ltd.     

From marine ecology to crime analysis: Improving the detection of serial sexual offences using a taxonomic similarity measure

Jaccard has been the choice similarity metric in ecology and forensic psychology for comparison of sites or offences, by species or behaviour. This paper applies a more powerful hierarchical measure—taxonomic similarity (Δ_s), recently developed in marine ecology—to the task of behaviourally linking serial crime. Forensic case linkage attempts to identify behaviourally similar offences committed by the same unknown perpetrator (called linked offences). Δ_s considers progressively higher‐level taxa, such that two sites show some similarity even without shared species. We apply this index by analysing 55 specific offence behaviours classified hierarchically. The behaviours are taken from 16 sexual offences by seven juveniles where each offender committed two or more offences. We demonstrate that both Jaccard and Δ_s show linked offences to be significantly more similar than unlinked offences. With up to 20% of the specific behaviours removed in simulations, Δ_s is equally or more effective at distinguishing linked offences than where Jaccard uses a full data set. Moreover, Δ_s retains significant difference between linked and unlinked pairs, with up to 50% of the specific behaviours removed. As police decision‐making often depends upon incomplete data, Δ_s has clear advantages and its application may extend to other crime types. Copyright © 2007 John Wiley & Sons, Ltd.     

Crime scene actions and offender characteristics in a sample of Finnish stranger rapes

The crime scene actions employed by offenders in stranger rapes were analysed in relation to offender characteristics. Data were drawn from an official police database and consisted of stranger rapes occurring in Finland between 1992 and 2001 (n = 100). The structure of dichotomous variables derived through a content analysis of crime scene actions and offender characteristics were analysed with non‐metric multidimensional scaling (MDS). The first analysis revealed three separate action themes, with thematic emphases on Hostility, Involvement or Theft. The MDS‐solution for offender characteristics suggested four themes: Conventional; Psychiatric/Elderly; Criminal/Violent; and Criminal/Property. Each case was assigned to one of the themes or as a hybrid in order to analyse the associations between action themes and characteristics. The only significant association was found between the action theme, Theft and characteristics theme Criminal/Property. The results are discussed in relation to previous research. Copyright © 2004 John Wiley & Sons, Ltd.     

Spatial patterns of Indian serial burglars with relevance to geographical profiling

Earlier studies in Western countries have shown remarkably consistent spatial patterns in serial offenders, mainly for serious crimes notably serial killing and rape, but also (although with less clear patterns) for burglary. The universality of such spatial patterns are of theoretical interest in contributing to our understanding of criminal spatial behaviour and have practical significance for the possibility of using geographic profiling in developing countries. As such, burglars in India provide a particularly interesting test of the generality of the observed spatial consistencies. Information was therefore obtained on the offence location choices of 30 burglars, committing 150 offences in the Rourkela and Keonjhar districts of India. The home to crime distances were compared with those from developed countries, revealing similar but slightly shorter distances. In addition, the domocentricity of criminal spatial activity, reflected in the ‘marauder’ model (Canter & Gregory 1994) was tested through examination of the Canter Circle hypothesis, the mean interpoint distances (as they related to average distances from home), and the home base ‘search costs’ using a geographical profiling system (Dragnet). Overall the study found that the spatial patterns of the sample of Indian Burglars were not very different from their counterparts in the UK, North America and Australia, showing that the areas in which an offender is active tend to be shaped by, and, relatedly, close to, where he or she lives, irrespective of the part of the world in which this is. These results suggest that geographical profiling systems such as Dragnet would be productively used on the Indian sub‐continent. The results also contribute to our understanding of possible universalities in offender spatial behaviour. Copyright © 2006 John Wiley & Sons, Ltd.     

Simple structure in component analysis techniques for mixtures of qualitative and quantitative variables

Several methods have been developed for the analysis of a mixture of qualitative and quantitative variables, and one, called PCAMIX, includes ordinary principal component analysis (PCA) and multiple correspondence analysis (MCA) as special cases. The present paper proposes several techniques for simple structure rotation of a PCAMIX solution based on the rotation of component scores and indicates how these can be viewed as generalizations of the simple structure methods for PCA. In addition, a recently developed technique for the analysis of mixtures of qualitative and quantitative variables, called INDOMIX, is shown to construct component scores (without rotational freedom) maximizing the quartimax criterion over all possible sets of component scores. A numerical example is used to illustrate the implication that when used for qualitative variables, INDOMIX provides axes that discriminate between the observation units better than do those generated from MCA.The Netherlands organization for scientific research (NWO) is gratefully acknowledged for funding this project. This research was conducted while the author was supported by a PSYCHON-grant (560-267-011) from this organization. The author is obliged to Jos ten Berge for his comments on an earlier version.