No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
A Quantitative Analysis of Behavioral Case Linkage for Practice
Abstract
Although many studies have been conducted on behavioral profiling of homicide offenders, none of these have empirically shown that this technique has scientific validity and solid accuracy. Based on data from 839 Japanese homicide cases, we constructed multivariate logistic regression models to infer offenders' characteristics of sex, age, stranger offender, criminal record, single offender, member of organized crime, and residential area. We evaluated validity with a tenfold cross-validation procedure and a receiver operating characteristic (ROC) analysis. The results indicated that the area under the ROC curve ranged from .70 to .87, suggesting moderate and sufficient accuracy. Such a multivariate approach would be useful to reduce costs and minimize errors in the early stages of an investigation, especially for jurisdictions that possess large homicide databases.
... This analysis of whether a series of crimes is committed by the same offender(s) is an offender profiling process. It has been called 'case linkage' ( Bennell and Jones, 2005;Fujita et al., 2011;Woodhams and Toye, 2007), 'linkage analysis' ( Hazelwood and Warren, 2003), 'linking cases' ( Canter and Youngs, 2009), and 'comparative case analysis' ( Bennell and Canter, 2002;Canter and Youngs, 2009). As described by Canter and Youngs (2009), the analyst first observes the crime scenes, collects information on the cases and arranges the data. ...
The effectiveness of stylometric analysis of Japanese communications from criminals for linking crimes through text mining was investigated. Forty letters of threats and blackmails in the ‘Glico–Morinaga’ case were analyzed as originals. These original letters were compared with four copycat letters in which the suspect in the ‘Kuroko’s Basketball’ (KB) case imitated the writings in the ‘Glico–Morinaga’ (GM) case. We focused on four stylometric features in the writings: (a) bigram of characters, (b) bigram of part-of-speech taggers, (c) unigram of vocabularies, and (d) frequency of use of Kanji, Hiragana, and Katakana. We analyzed these stylometric features using correspondence analysis and hierarchical cluster analysis. The results showed that all the ‘KB’ texts gathered in the same crowd or cluster, which was separate from those including the ‘GM’ texts. Moreover, the ‘GM’ texts were not consistently distant from each other. This study suggested that stylometric analysis could distinguish between original and copycat communications, and appropriately link crimes on the basis of written information.
The effectiveness of identifying the author of an illegal document by using text mining was investigated. The suspected writing evaluated in this study was a claim of responsibility written by a 14-year-old boy, which stated that he committed the “Kobe child murders” in 1997. It was compared with control writings including confessions, and an essay that we knew were written by the same boy, as well as with irrelevant materials including various essays written by five junior high school students, and claims of responsibility in four past criminal cases. First, the writings in each document were digitalized and converted to text files. Then, the relative frequencies of bigram of letters, bigram of part-of-speech taggers, sentence lengths of each document, and rate of using Kanji, Hiragana, and Katakana were calculated. Results of sammon multi-dimensional scaling and hierarchical cluster analysis indicated that the text in the suspected writing was arranged identically or similarly to groups of texts in control materials, where they were arranged differently from groups of texts in irrelevant materials. In a separate analysis, the suspected writing was substituted with a document written by a different offender and we conducted the identical procedure described above. Results demonstrated that texts in the suspected writing were in a different form control and irrelevant texts. These results indicated the effectiveness of identifying an author by using text mining when examining forensic documents.
The extrapolation of characteristics of criminals from information about their crimes, as an aid to police investigation, is the essence of ‘profiling'. This paper proposes that for such extrapolations to be more than educated guesses they must be based upon knowledge of (1) coherent consistencies in criminal behaviour and (2) the relationship those behavioural consistencies have to aspects of an offender available to the police in an investigation. Hypotheses concerning behavioural consistencies are drawn from the diverse literature on sexual offences and a study is described of 66 sexual assaults committed by 27 offenders against strangers. Multivariate statistical analyses of these assaults support a five-component system of rapist behaviour, reflecting modes of interaction with the victim as a sexual object. The potential this provides for an eclectic theoretical basis to offender profiling is discussed.
The newly emerging area of Investigative Psychology provides a behavioural science basis for crime detection by examining investigative processes and criminal behaviour. It draws upon a range of material collected by law enforcement agencies that is not widely utilised in the social sciences. This may be regarded as a form of non- reactive, unobtrusive data that has many of the advantages originally promoted by Webb, Campbell, Schwartz and Sechrest (1966) and more recently explored by Lee (2000). The value of such data, derived from police sources, has been demonstrated in a variety of Investigative Psychology studies. However, law enforcement material is not usually collected as data but rather as evidence. Consideration is therefore given to how to address the challenges this poses. The unobtrusive measures derived from police investigations provide a different perspective on crime and other aspects of human actions from that based on more conventional sources of data such as questionnaires and interviews. To assist in the effective use of measures derived from police information a framework for considering this material is proposed reflecting the range of sources of measures that Lee (2000) identified; personal records, running records, physical traces, and simple observation. As in other areas, close attention to the methods of collecting such material can considerably improve its utility. The measures being utilized in Investigative Psychology therefore offer some fruitful directions for other areas of social science research. Development of these measures can also improve the effectiveness of
This article explores how unobtrusive research methods popularized by Webb et al. (1966) can be utilized in forensic research. In particular, the value of the approach is considered with special reference to examining investigative processes and criminal behaviour. Webb et al.’s three non-reactive types of unobtrusive measures (physical traces, archival material and simple observation) are discussed in relation to a variety of types of material collected across the course of police enquiries. The breadth and variation of this material is emphasized with special consideration of its utility for research. We illustrate how the limitations and benefits identified by Webb et al. are echoed within the forensic domain and how close attention to the methods of collecting such material developed by researchers may improve the effectiveness of investigations.
It is hypothesized that there will be behavioural consistencies in the actions of arsonists when committing a crime that characterize them. The themes underlying these observable differences can be used to help us understand the nature of the offence. With arson, one such observable difference is hypothesized to relate to the target or focus of the attack. The study tested whether consistencies could be found that distinguish person‐oriented from object‐oriented arsons. A second proposed facet of arson actions relates to the motivational category underlying the act, being either instrumental or expressive. It was also hypothesized, therefore, that there would be a distinction in the arson actions between fires set for a clear instrumental purpose, and those which may be regarded as emotional acting‐out. The hypothesis that these four themes would differentiate arsonists was tested by analysing 175 solved arson cases from across England. The case files were content analysed to produce 42 behavioural variables taken from both the crime reports and witness statements. In order to test the hypotheses of differentiation a smallest space analysis was carried out.
The results support this framework giving rise to four distinct themes to arson from which scales with reasonable alpha scores could be derived. Two relate to expressive acts, ( a ) those that are realized within the arsonist's own feelings, being analogous to suicide, and ( b ) those that are acted on objects, like the burning of symbolic buildings. The other two relate to instrumental acts, ( c ) those that are for personal indulgence, similar to personal revenge, and ( d ) those that have an object focus such as hiding evidence from a crime. A further test of the validity of these four themes was to examine the typical characteristics of the people who committed the different types of arson. Four scales of arsonists' characteristics were developed. These were found to have appropriate, statistically significant correlations with the four themes.
The implications of these findings for understanding the varieties of arson as revealed through the actions that occur are discussed, as well as the implications for arson investigations. It is speculated that this framework may provide a general model for considering a wide range of crimes.
The purpose of this study is to determine if readily available information about commercial and residential serial burglaries, in the form of the offender's modus operandi, provides a statistically significant basis for accurately linking crimes committed by the same offender. Logistic regression analysis is applied to examine the degree to which various linking features can be used to discriminate between linked and unlinked burglaries. Receiver operating characteristic (ROC) analysis is then performed to calibrate the validity of these features and to identify optimal decision thresholds for linking purposes. Contrary to crime scene behaviours traditionally examined to link serial burglaries, the distance between crime site locations demonstrated significantly greater effectiveness as a linking feature for both commercial and residential burglaries. Specifically, shorter distances between crimes signalled an increased likelihood that burglaries were linked. Thus, these results indicate that, if one examines suitable behavioural domains, high levels of stability and distinctiveness exist in the actions of serial burglars, and these actions can be used to accurately link crimes committed by the same offender. Copyright © 2005 John Wiley & Sons, Ltd.
Case linkage, the identification of crimes suspected of being committed by the same perpetrator on the basis of behavioral similarity, and offender profiling, the inference of offender characteristics from offense behaviors, are used to advise police investigations and, in relation to case linkage, have been admitted in legal proceedings. Criteria for expert evidence, such as the Daubert criteria (Daubert v. Merrell Dow Pharmaceuticals, 1993), place stringent conditions on the admissibility of expert evidence. The future contribution of these practices to legal proceedings depends, in part, on whether they are underpinned by hypotheses that are testable and supported. The 3 hypotheses of offender behavioral consistency, of offender behavioral distinctiveness, and of a homology (direct relationship) between offender characteristics and behavior were empirically examined using a sample of serial commercial robberies. Support was found for the former 2 hypotheses but not for the last. The findings of the 2 studies have implications for the future development of these practices, for legal practitioners evaluating expert evidence, and for the implementation of public policy.
Terminal restriction fragment length polymorphism (T-RFLP) is a culture-independent method of obtaining a genetic fingerprint of the composition of a microbial community. Comparisons of the utility of different methods of (i) including peaks, (ii) computing the difference (or distance) between profiles, and (iii) performing statistical analysis were made by using replicated profiles of eubacterial communities. These samples included soil collected from three regions of the United States, soil fractions derived from three agronomic field treatments, soil samples taken from within one meter of each other in an alfalfa field, and replicate laboratory bioreactors. Cluster analysis by Ward's method and by the unweighted-pair group method using arithmetic averages (UPGMA) were compared. Ward's method was more effective at differentiating major groups within sets of profiles; UPGMA had a slightly reduced error rate in clustering of replicate profiles and was more sensitive to outliers. Most replicate profiles were clustered together when relative peak height or Hellinger-transformed peak height was used, in contrast to raw peak height. Redundancy analysis was more effective than cluster analysis at detecting differences between similar samples. Redundancy analysis using Hellinger distance was more sensitive than that using Euclidean distance between relative peak height profiles. Analysis of Jaccard distance between profiles, which considers only the presence or absence of a terminal restriction fragment, was the most sensitive in redundancy analysis, and was equally sensitive in cluster analysis, if all profiles had cumulative peak heights greater than 10,000 fluorescence units. It is concluded that T-RFLP is a sensitive method of differentiating between microbial communities when the optimal statistical method is used for the situation at hand. It is recommended that hypothesis testing be performed by redundancy analysis of Hellinger-transformed data and that exploratory data analysis be performed by cluster analysis using Ward's method to find natural groups or by UPGMA to identify potential outliers. Analyses can also be based on Jaccard distance if all profiles have cumulative peak heights greater than 10,000 fluorescence units.
It is hypothesized that stranger rape victim statements will reveal a scale of violation experienced by the victim, ranging from personal violation, through to physical violation, and finally, at the most extreme level, sexual violation. It is also hypothesized that offences can be differentiated in terms of one of four themes: hostile, controlling, stealing, or involving. To test these hypotheses, crime scene data from 112 rapes were analyzed by the multi-dimensional scaling procedure Smallest Space Analysis. The results provide empirical support for a composite model of rape consisting of four behavioral themes as different expressions of various intensities of violation. The results also suggest that stranger rapes may be less about power and control than about hostility and pseudo-intimacy. The proposed model has implications for the classification of rape, the investigation of sexual crimes, and the treatment of victims.
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.
The origins of ‘Offender Profiling’ in the advice given by police medical advisors and other experts to criminal investigations are briefly outlined. The spread of such advice to police enquiries across the United States in the early 1970s, culminating in its uptake by Special Agents of the FBI in the mid‐1970s and the widespread promotion of their services through the fictional writings of Thomas Harris and others is noted.
The development beyond the early application to serial killer investigations, and the focus on psychopathological explanations, to cover the full gamut of crime from, for instance, arson and burglary to terrorism, is briefly reviewed. The consideration of the social psychological processes inherent in criminality as well as the characteristics of individual offenders also broadens out the concerns of the field. The linking of crimes to a common offender as well as predicting their future actions further widens the range of issues to be dealt with.
The many psychological and practical questions raised by these ‘profiling’ activities are summarised. These include questions of inference and prediction, about criminals and their crimes, both about their characteristics and about the spatial patterns of their activities. Related topics concerning the sources of information for both investigators and research are also summarised. These cover the full range from interviewing witnesses to the management of informants. The complexity of information management and inference derivation points to the need to understand investigative decision‐making and how it can be supported.
These other issues, beyond those inherent in ‘profiling’, such as data integrity and investigative decision support, taken with the central ‘profiling’ questions leads to the identification of a new domain of applied psychology, ‘Investigative Psychology’. It is argued that the core topic of this domain, as in any emerging science, is how to appropriately describe and classify the central matters under consideration, i.e. criminals and their activities. The difficulties in setting up reliable, robust and valid classification schemes are discussed and approaches to overcoming these difficulties considered. It is emphasised that although many researchers have found Multi‐Dimensional Scaling procedures to be productive they are only one of many fruitful sets of approaches that are possible.
The increasing variety of areas, for which Investigative Psychology is relevant, from tax evasion to peace keeping, and from evidence in court to organisational threat management, is briefly reviewed. In conclusion it is noted that Investigative Psychology can be considered as a general approach to problem solving relevant far beyond criminal investigations. This new Journal of Investigative Psychology and Offender Profiling therefore has rich and wide‐ranging potential. Copyright © 2004 John Wiley & Sons, Ltd.
The current study examines the performance of cluster analysis with dichotomous data using distance measures based on response pattern similarity. In many contexts, such as educational and psychological testing, cluster analysis is a useful means for exploring datasets and identifying un-derlying groups among individuals. However, standard approaches to cluster analysis assume that the variables used to group observations are continu-ous in nature. This paper focuses on four methods for calculating distance between individuals using dichotomous data, and the subsequent introduc-tion of these distances to a clustering algorithm such as Ward's. The four methods in question, are potentially useful for practitioners because they are relatively easy to carry out using standard statistical software such as SAS and SPSS, and have been shown to have potential for correctly grouping ob-servations based on dichotomous data. Results of both a simulation study and application to a set of binary survey responses show that three of the four measures behave similarly, and can yield correct cluster recovery rates of between 60% and 90%. Furthermore, these methods were found to work better, in nearly all cases, than using the raw data with Ward's clustering algorithm.
This paper presents a general statistical methodology for the analysis of multivariate categorical data arising from observer reliability studies. The procedure essentially involves the construction of functions of the observed proportions which are directed at the extent to which the observers agree among themselves and the construction of test statistics for hypotheses involving these functions. Tests for interobserver bias are presented in terms of first-order marginal homogeneity and measures of interobserver agreement are developed as generalized kappa-type statistics. These procedures are illustrated with a clinical diagnosis example from the epidemiological literature.
Two females, Denean Worms and Brenda Hughes, were murdered in separate events in Cranbrook, British Columbia in 1984 within three months of each other. Terrence Wayne Burlingham was found guilty of both murders and he appealed. The Supreme Court of Canada granted Burlingham a new trial in the Worms case, but no evidence from Burlingham's confession nor the murder weapon could be used. The Crown counsel requested an evaluation of the two murders to determine if they were committed by the same person. The analyses of those murders revealed that they were linked by a personal "signature" of the killer. The murder cases reported here demonstrate a control-oriented signature. The killer used a .410 shotgun as his method of control and death, engaged in overkill of each victim by shooting them twice in the head, and left the victims in sexually degrading positions. Another signature feature was the absence of typical wounds to the victims which would be expected from a serial sex offender. All of these characteristics, in combination, accounted for this killer's personal expression.
The authors developed a data-based profiling system in order to support offender profiling. The system stored incident records of prior offenders. Inputting offence details of an unsolved incident, a probability score was assigned to each prior offender in the system; the score represented the behavioral similarity with the unsolved incident. The system then ranked all offenders in the system according to the probability scores, and prioritized the high-ranked offenders as possible suspects. Moreover, the system inferred the characteristics of unknown offenders by accumulating characteristics of the high-ranked offenders. The system achieved promising accuracy, especially for linking crimes to perpetrators. In 45 out of 81 simulation trials, the target offenders were retrieved as a rank score of 1 from among 868 sex offenders.
Unobtrusive measures
- E J Webb
- D T Campbell
- R D Schwartz
- L Sechrest