Multidimensional scaling and the analysis of sexual offence behaviour: a reply to Sturidsson et al.

Abstract

Sturidsson et al. (Psychology, Crime & Law, 12, 221–230, 2006) published an article on the replicability of early multidimensional scaling (MDS) behavioural research by Canter and Heritage (Journal of Forensic Psychiatry, 1, 185–212, 1990). Sturidsson et al. were unable to find a ‘readily apparent, simple interpretation’ of their MDS solution and therefore suggested that the MDS approach may be flawed as a device for interpreting sexual assault. The current paper explains how the methods employed by Sturidsson et al. represent an incorrect use of MDS, specifically in their input of Jaccard similarity data into the MDS (ALSCAL) procedure in SPSS as opposed to the correct MDS (PROXSCAL) procedure. A comparative example is used to illustrate how adopting the ALSCAL procedure will result in an inverse MDS solution, where objects appearing closer together are more dissimilar, making interpretation of the MDS plot meaningless in this context. The current paper discusses this methodological flaw in greater depth and in the context of behavioural investigative research using MDS in general. It concludes with some general guidelines in utilizing the MDS procedure in SPSS specific to criminal profiling research.     

In defence of multidimensional scaling for the analysis of sexual offence behaviour: cautionary notes regarding analysis and interpretation

Abstract

Multidimensional scaling (MDS) has become a popular form of statistical analysis in the field of psychology and law. A recent paper in this journal by Sturidsson and colleagues (12, 221–230, 2006) used MDS to replicate an earlier study of sexual offence behaviour by Canter and Heritage (Journal of Forensic Psychiatry, 1, 185–212, 1990). They concluded that the model was not replicable and cast grave doubts on the applicability of MDS for the analysis of sexual offences. In this paper I argue that neither conclusion is correct. Initial difficulties in interpretation of the MDS results were clearly due to errors in calculation that should have been recognised. Indeed, a frequency analysis of Sturidsson and colleagues’ variables indicated that their results were in fact remarkably congruent with those of Canter and Heritage. It is concluded that statistical software should not obviate the need for a working knowledge of the statistics being computed.     

Crime scene actions and offender characteristics in arsons

Purpose. Crime scene actions displayed by offenders in arsons were analysed in relation to offender characteristics and self‐reported motives. Methods. Data were drawn from an official police database and consisted of a random sample of arsons occurring in Finland between 1990 and 2001 (N = 189). The structure of dichotomous variables derived through a content analysis of crime scene actions and offender characteristics was analysed using non‐metric multidimensional scaling (MDS) analyses. Results. The analyses revealed four separate crime scene action types, with thematic emphasis on either expressive or instrumental motivation and the target being either a person or object. The MDS solution for offender characteristics suggested four types of offenders: Self‐destructive, Serial, Criminal, and Adolescent. In order to analyse the associations between action and characteristics themes, each case was assigned to one of the action and characteristic themes or a hybrid based on the frequency of variables present from each theme. Several statistically significant associations between the action and characteristic themes were found. Self‐reported motives were also analysed in relation to the themes. Conclusions. The results support the thematic classification of arsonists presented in Canter and Fritzon (1998). The results are discussed with regard to earlier research and deducing offender characteristics based on crime scene actions.     

Juvenile firesetters: Crime scene actions and offender characteristics

Purpose. To investigate whether and to what extent the thematic structure of crime scene actions in arsons identified in Canter and Fritzon (1998) is replicated for juvenile firesetters and to explore whether any associations between the crime scene action themes and offender characteristics would be evident. Methods. The crime scene actions and offender characteristics of 61 male and 5 female juvenile firesetters (aged 6‐17 years) were examined. The data were drawn from a larger database originally collected and content analysed in Fritzon (1998). In total, 43 dichotomous crime scene actions, 17 offender background characteristics and offender criminal record variables had been coded. Smallest space analysis was employed to examine the configuration of crime scene action and offender characteristic variables. The associations between the crime scene actions and offender characteristics, as well as the criminal record variables, were analysed using the correlation. Results. Distinct structural themes for crime scene actions were found in juvenile firesetting, similar to those identified in Canter and Fritzon (1998). Contrary to Canter and Fritzon, only two groups of background characteristics were identified, depressed and delinquent, the latter being more common and related to an instrumental form of firesetting. The expressive form of firesetting was associated with offenders' psychopathology and female gender. The presence of a crime scene action theme was associated with the offender's age. Conclusions. The structural themes of firesetting behaviour appear to transpire early. The background characteristics of juvenile firesetters indicate that juvenile firesetting is often associated with antisocial behaviour and psychopathology, deserving, therefore, disparate prevention, intervention and investigation programmes.     

Investigating the Reliability of the Violent Crime Linkage Analysis System (ViCLAS) Crime Report

The Violent Crime Linkage Analysis System (ViCLAS) is a computerized investigative aid, developed by the Royal Canadian Mounted Police, which is used to increase information sharing and enable crime linkages in serious crime investigations. The current study examines the reliability of the ViCLAS Crime Report by calculating the level of agreement among two samples of police officers who received one of two fictitious crime scenarios; homicide (n = 116) or sexual assault (n = 121). Results from this study show that the observed inter-rater reliability among police officers completing the ViCLAS booklet was 79.30% for the homicide scenario, and 87.70% for the sexual assault scenario. Implications of the results are discussed.     

Inter-rater reliability of dynamic risk assessment with sexual offenders

Abstract

This paper describes the inter-rater reliability of the Structured Assessment of Risk and Need (SARN, formerly known as Structured Risk Assessment). The SARN is a structured framework for identifying sexual offenders’ dynamic risk factors. The SARN comprises 16 dynamic risk factors, categorized into four domains: Sexual Interests, Distorted Attitudes, Socio-Affective Functioning and Self-Management. Two studies, utilizing three samples, are reported. Study 1 examined the inter-rater reliability of four SARN cases with a sample of seven expert raters. Results indicated high inter-rater reliability amongst these participants. Study 2 examined the reliability of SARN with two samples who had received training before supplying inter-rater data (N=88). Results provided some support for the reliability of SARN. However, strength of reliability was dependent upon the method of analysis applied (percentage agreement, Cohen's Kappa, intra-class correlation coefficients). These results are discussed in terms of their clinical and methodological implications.     

Group size does matter: differences among sexual assaults committed by lone,double, and groups of three or more perpetrators

ABSTRACT

We compared a total of 340 sexual assault cases based on offender characteristics, victim characteristics, and offence behaviours to identify differences among sexual assaults committed by lone, double, and groups of three or more perpetrators in South Korea. Several differences were found among sexual assaults committed by varying numbers of offenders. First, lone, double, and groups of three or more perpetrators differed with respect to age, marital status, and employment status. Second, substantial differences in victim characteristics were observed with respect to age and the victim–offender (v–o) relationship. Third, significant differences in offence behaviours, including offence type, planning, method of approach, knife possession, and stealing property, were found. Furthermore, we identified several variables that significantly predict group size using logistic regression analyses. The study has theoretical implications for understanding how group dynamics influence the behaviours of rape perpetrators, and practical implications for victim treatment and offender intervention.     

The development and initial testing of the AIM2 framework to assess risk and strengths in young people who sexually offend

Abstract

This paper describes the AIM2 assessment framework and the process of its development and initial testing. AIM2 is used to assess areas of concerns and strengths of young people. Some preliminary analysis is described, including the correlation of assessment items, their ability to discriminate between cases, their inter-rater reliability and a small-scale recidivist study. These analyses were used to develop AIM2. Results from the recidivist study identified a particular set of concern and strength factors associated with recidivism in the sample. We conclude that the inclusion of strengths-related items are important to consider as protective factors when assessing the risk of further sexually abusive behaviours. Limitations and future directions of AIM2 are also discussed.     

Cold Case Homicides: DNA Testing of Retained Autopsy Sexual Assault Smears

Archival medical examiner specimens may contain perpetrator DNA evidence useful in unsolved (“cold case”) homicides. The Office of the Chief Medical Examiner (OCME) histology slide archives were searched for sexual assault smears for all 376 female homicides from 1990 to 1999. Of these, the OCME had sexual assault smears on 84 of which 13 slides had sperm. Of these 13, six were still unsolved. DNA profiles were obtained on all six (5 from smears and one from swabs). Combined DNA Index System ( submission resulted in two matches (“hits”) for new suspects. In addition, three suspects were eliminated in two cases. Our review of archival sexual assault smears resulted in DNA profiles that were able to assist in the investigation of four cold case homicide investigations. It may be worthwhile for medical examiner offices to search their archival histology slides for sexual assault smears on previously unsolved cases particularly those prior to the mid‐1990s when DNA testing was less widely available.     

Background Levels of Salivary-α-amylase Plus Foreign DNA in Cases of Oral Intercourse: a Female Perspective

Saliva plus DNA from a suspect is commonly encountered in sexual assault cases on bodily swabs. However, without background knowledge, the weight of this evidence is unknown. It may indicate the presence of saliva resulting from cunnilingus, or it may represent indirect transfer. In this study, females who refrained from cunnilingus donated 43 items of underwear and 19 vaginal swabs. The samples were subjected to Phadebas^®, RSID^™-Saliva and mRNA profiling and were subsequently DNA-profiled to determine the prevalence of background saliva in the female population. The results report that 15.8% of females who refrained from cunnilingus were positive for saliva and a further 10.5% also had DNA from unknown source(s). These findings of the rate of indirect transfer were evaluated with the Bayesian approach, and it was found that the evidence of saliva plus a high foreign DNA source adds moderately strong support to the allegation of cunnilingus.     

Examining the relationship between sexual offenders and their victims: Interpersonal differences between stranger and non-stranger sexual offences

Abstract

The present study examined the behavioural differences in sexual assault offences in relation to the offender–victim relationship (stranger versus non-stranger). These differences were examined specifically in the context of four interpersonal themes of interaction: dominance, submission, hostility and cooperation. The details of 100 sexual offence cases (50 stranger and 50 non-stranger) were content-analysed, generating 58 dichotomous variables, covering offender and victim behaviour during the offence. χ² tests comparing the two samples found that offenders who were strangers to their victims were more likely than non-stranger offenders to display behaviours that indicate a hostile, violent offence style. In contrast, those offenders who knew their victims were more likely than strangers to display a less violent and more personal, compliance-gaining offence style. These findings are discussed in terms of their implications for offender rehabilitation and victim support.     

Applicability of the ParaDNA® Screening System to Seminal Samples

Seminal fluid represents a common biological material recovered from sexual assault crime scenes. Such samples can be prescreened using different techniques to determine cell type and relative amount before submitting for full STR profiling. The ParaDNA^® Screening System is a novel forensic test which identifies the presence of DNA through amplification and detection of two common STR loci (D16S539 and TH01) and the Amelogenin marker. The detection of the Y allele in samples could provide a useful tool in the triage and submission of sexual assault samples by enforcement authorities. Male template material was detected on a range of common sexual assault evidence items including cotton pillow cases, condoms, swab heads and glass surfaces and shows a detection limit of 1 in 1000 dilution of neat semen. These data indicate this technology has the potential to be a useful tool for the detection of male donor DNA in sexual assault casework.     

犯罪剖绘:科学还是艺术?

犯罪剖绘(criminal profiling)为分析(连续)犯罪行为型态的技术,藉由统计技术分析潜在犯罪者、被害者与犯罪现场迹证等三者之间相互联结可能性。然犯罪剖绘技术应用至案件连结分析,可根据犯罪现场所遗留迹证(含心理迹证),预测/推论犯罪者之人口背景特征,并将这些案件之犯罪现场迹证系统化处理,建构成数据库,有助警察侦办连续型犯罪。经由案件连结分析所得证据,能否成为法庭上论罪之依据,在英美国家必须通过Frye法则与Daubert测试法则检验,即是否已经通过科学性检验,或同一学术领域同侪认可验证过程。因此,后续研究可朝更精确统计分析方法迈进,以提升案件连结分析数据库对于预测未发现犯罪者之准确率及其法庭证据力。     

Evaluation of usefulness of further Y-STR analysis in sexual assault cases on PSA positive samples resulting in female autosomal STR profiling

Samples from male to female sexual assault cases that are positive for the presumptive prostate specific antigen (PSA) test often do not result in a male autosomal STR profile. Due to highly unequal proportions of female and male DNA in typical sexual assault samples, routine autosomal STR analysis often fails to detect the DNA of the assailant, even after differential extraction of the samples. Previous studies have already shown the value of Y-STR analysis in such cases [1]. In Belgium, forensic DNA laboratories are only allowed to perform Y-STR profiling on sexual assault samples by a specific requisition, after routine autosomal STR analysis has been performed. However, a request for additional Y-STR analysis is rather exceptional.In this study, we evaluated the usefulness of further Y-STR analysis. For 100 PSA positive rinses and swabs from male to female sexual assault cases resulting in female autosomal STR profiling, 7% resulted in a full or partial Y-STR profile useful for comparison, using the 23-loci Y-STR PowerPlex Y23 System (Promega). The success rate raised to 12.5% with a higher DNA input in the PCR mix. In conclusion, these results support the usefulness of performing Y-STRs analysis on the sperm DNA extracts to identify the alleged assailant in sexual assault cases.     

Differentiation of Twenty‐One Glitter Lip‐Glosses by Pyrolysis Gas Chromatography/Mass Spectroscopy*

Abstract: Differentiation of 21 glitter lip‐glosses from seven manufacturers was attempted by pyrolysis gas chromatography/mass spectroscopy. Samples were pyrolyzed on a ribbon probe at 800°C for 20 sec and analyzed with an Agilent^® 6890N Network GC System and Agilent^® 5973 Network Mass Selective Detector with MSD Productivity ChemStation^® Data Analysis software. The total ion chromatograms obtained were examined and differences in the presence or absence of certain chromatographic peaks corresponding to certain pyrolysis products (e.g., styrene, cyclohexane) noted. In cases where the total ion chromatograms between lip‐glosses were similar, select ion profiling was performed. Of the 21 lip‐glosses, 15 were differentiated by either the total ion chromatograms alone or through select ion profiling. Considering that lip‐glosses are typically worn by young women (who are disproportionately victims of sexual assault), this study offers the potential of being able to provide investigative leads in sexual assault investigations with evidentiary samples of this kind.     

Using Bayes’ theorem in behavioural crime linking of serial homicide

Purpose. The study extends research by Santtila et al. (2008) by investigating the effectiveness of linking cases of serial homicide using behavioural patterns of offenders, analysed through Bayesian reasoning. The study also investigates the informative value of individual behavioural variables in the linking process. Methods. Offender behaviour was coded from official documents relating to 116 solved homicide cases belonging to 19 separate series. The basis of the linkage analyses was 92 behaviours coded as present or absent in the case based on investigator observations on the crime scene. We developed a Bayesian method for linking crime cases and judged its accuracy using cross‐validation. We explored the information added by individual behavioural variables, first, by testing if the variable represented purely noise with respect to classification, and second, by excluding variables from the original model, one by one, by choosing the behaviour that had the smallest effect on classification accuracy. Results. The model achieved a classification accuracy of 83.6% whereas chance expectancy was 5.3%. In simulated scenarios of only one and two known cases in a series, the accuracy was 59.0 and 69.2%, respectively. No behavioural variable represented pure noise but the same level of accuracy was achieved by analysing a set of 15, as analysing all 92 variables. Conclusion. The study illustrates the utility of analysing individual behavioural variables through Bayesian reasoning for crime linking. Feasible applied use of the approach is illustrated by the effectiveness of analysing a small set of carefully chosen variables.     

The role of sexual interests and situational factors on rapists' modus operandi: Implications for offender profiling

Purpose. Although it has often been suggested that there is a direct relationship between an individual's sexual interests and behaviours exhibited during the crime they commit, few studies have investigated this question empirically. The purpose of the present study was thus to examine the role of sexual interests and situational factors as to their possible relationship to three components of rapists' modus operandi: (a) the level of organization of the offence, (b) the level of force used by the offender, and (c) the level of injury inflicted on the victim during the sexual assault. Methods. This study is based on a sample of 118 offenders who sexually assaulted a female aged 16 or over. All participants were assessed phallometrically and through the CQSA, a computerized questionnaire. Data were analysed using multiple regression analyses. Results. Our findings showed links between sexual interests, situational factors, and rapists' modus operandi. Firstly, individuals demonstrating a greater sexual interest in nonsexual violence showed a higher level of organization in the modus operandi. Secondly, alcohol consumption prior to the offence was related to a higher level of coercion. Finally, a negative emotional state prior to the crime was related to a high level of injury inflicted on the victim. Conclusions. Despite the fact that several authors postulated a direct link between the offender's sexual interests and his behaviour at the crime scene, our results only partially support this hypothesis. Moreover, our results partly support the fact that crime scene behaviour is related to offenders' personal characteristics, challenging an assumption of criminal profiling. We still believe that the modus operandi is related to offenders' personal attributes. It is, however, dynamic and may fluctuate due to certain situational factors related to offenders and victims. Future studies should take into account situational factors related to offenders and their victims.     

Victim's perceptions of initial police responses to robbery and aggravated assault: Does race matter?

Despite the fact that police officers are usually the first persons within the criminal justice system to respond to a criminal victimization, the majority of research investigating racial discrimination within the system has examined primarily the effects of race on adjudication outcomes which occur after initial police interventions, such as conviction decisions and sentences. Very little empirical effort has been devoted to examining the effects of race on early police responses to a reported victimization. Using data from the National Crime Victimization Survey from 1987 to 1992, this paper investigates the effects of both the victim's and the offender's race on three police responses to robbery and aggravated assault: (1) police response time to the scene, (2) effort exerted by the police at the scene, and (3) likelihood of arrest. It was found that police were quicker to respond and also exerted more effort at the scene such as searching and taking evidence to incidents of black on white robbery compared to all other racial dyads. This relationship held even after controlling for other factors such as victim-offender relationship, poverty, injury to the victim, and victim's gender. No significant effects of race, however, were found when predicting the probability of arrest in cases of robbery. The effects of race on police responses to aggravated assault were more complicated. For assaults involving strangers, police were significantly more likely to exert additional effort at the scene if the victim was white and the offender was perceived to be black. This effect was reversed, however, for nonstranger assault victimizations. Police were significantlyless likely to exert effort at the scene or to make an arrest in black on white assaults involving nonstrangers. The most consistent predictors of arrest in both stranger and nonstranger assault victimizations were police response time, injury to the victim, and the incident occurring in a public setting.