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.     

Frequency of serial sexual homicide victimization in Virginia for a ten-year period

The frequency of serial sexual murder has been widely discussed, and estimates of the number of victims in the United States range from 500 to 6000 per year. This study attempted to quantify the number of serial sexual murder victims in Virginia for a ten-year period. Multiple sources of data were utilized, including Federal Bureau of Investigation (FBI) case files, FBI's Violent Criminal Apprehension Program (ViCAP) reports, Virginia State Police ViCAP reports, Virginia Homicide Investigators Association, and the Virginia Division of Forensic Science DNA database, to effectively cull out all the known serial sexual murder victims for the given time period. Review of these records revealed a total of 28 victims of serial sexual murder, compared with a total of 5183 murder victims for the same ten-year period. The frequency of serial sexual victimization was 0.5% of all homicides for the given period. These results highlight the unusual frequency of serial sexual murder.     

Identification of serial homicide victims in the "Green River Murder" investigation

The Green River Murder Investigation in King County, Washington, is currently the longest and most active serial murder investigation in U.S. history. To date, little information has been reported on methods used in identification of serial murder victims. In this paper, various methods used in victim identification are reviewed and difficulties encountered during the course of the investigation are described. The experience of the authors is presented in order to acquaint other agencies with problems of victim identification associated with these serial murder victims and to provide key methods that may be useful in other such investigations.     

Aggression homicide and rejection homicide: a communicative classification of homicide

Based on a 10-year sample of homicides (n = 50), the hypothesis was tested that it is possible to differentiate between aggression and rejection homicide. The aggression homicide results from the offender/victim relationship, which is no longer accepted for some reason. In contrast, in the rejection homicide the offender radically strives for a goal which can only be reached if the victim is eliminated. Based on forensic-psychiatric expert opinions (n = 50), the case analyses yielded 31 aggression homicides and 18 rejection homicides, one case could not be classified. Aggression homicides differed significantly from the rejection homicides with regard to their main motives. Hate in quarrel (n = 8), violent occupation of the victim (n = 7), delusions (n = 5), revenge (n = 3), self-defence (n = 2), and jealousy (n = 1) characterized the aggression homicides, whereas rejection homicides were dominated by economic motives (n = 14). Two offenders intended to get rid of the victim and one offender wanted to rescue a third person. From a forensic-psychiatric point of view, the pertinent statistical data (social data, diagnosis, criminal responsibility) are reported and the significance of the differentiation between aggression homicide and rejection homicide for medico-legal or criminological case profiling and for the prognosis of the risk potential is discussed.     

Piquerism in homicide: A knife wound analysis of a sexually motivated serial killer

As part of a wider research project into paraphilic disorders and homicide, a case is presented of a piquerist who achieved sexual satisfaction through the infliction of stabbing wounds. Against a background of sexual domination with willing and unwilling sexual partners, the offender stabbed four men to death. A study of the resulting autopsy reports, in association with the offender's testimony and crime scene documentation, was undertaken to discern a piquerist crime scene signature. An excessive number of stabbing wounds, the targeting of sexual areas, and a variation in wound size and depth were found to be evidence of piquerism. Although there was opportunity to commit a sexual assault upon victims, the preference was to penetrate them with a knife. It is contended that, even without any evidence of sexual assault at a crime scene, a knife wound analysis can reveal a sexual motivation in some cases of homicide.     

The case of Jeffrey Dahmer: sexual serial homicide from a neuropsychiatric developmental perspective

Sexual serial homicidal behavior has received considerable attention during the last three decades. Substantial progress has been made in the development of methods aimed at identifying and apprehending individuals who exhibit these behaviors. In spite of these advances, the origins of sexual serial killing behavior remain for the most part unknown. In this article we propose a biopsychosocial psychiatric model for understanding the origins of sexual serial homicidal behavior from both neuropsychiatric and developmental perspectives, using the case of convicted serial killer Jeffrey Dahmer as the focal point. We propose that his homicidal behavior was intrinsically associated with autistic spectrum psychopathology, specifically Asperger's disorder. The relationship of Asperger's disorder to other psychopathology and to his homicidal behavior is explored. We discuss potential implications of the proposed model for the future study of the causes of sexual serial homicidal crime.     

Luis Alfredo Garavito Cubillos: criminal and legal aspects of serial homicide with over 200 victims

This is the first scientific report on the crimes of the homosexual paedophile sadist Luis Alfredo Garavito Cubillos, based on a research stay of the authors in Columbia, and including discussions with the investigators, and the offender. Between 1992 and 1999, Garavito killed more than 200 children in the core age span between 8 and 13 years (as an exception, 6 to 16 years). His modus operandi remained stable. During daytime, he lured children of a lower social status out of crowded parts of the city into hidden areas that were overgrown with high plants. Garavito promised either payment for easy work, or drugs, or made other socially believable offers. The children were tied up, tortured, raped, and killed by at least one cut in the lateral part of the neck, or by decapitation. During the killings, Garavito was drunk. Even after his arrest (for attempted sexual abuse under a wrong identity) it was not immediately possible to track his crimes since Garavito had frequently changed his places of stay and his jobs. He also grew different hairdos and used wrong names. During his still ongoing confessions, he directs the investigators correctly to all scenes of crime spread over large parts of Columbia. In our report, we give an overview over the course of investigations, hint to similarities in the cases of the German serial killer Denke (1920's) and homosexual paedophile serial killer Jürgen Bartsch (1960's), and give preliminary impressions on the offender's personality. Furthermore, the violent environment and juridical peculiarities in Columbia are discussed. In spite of a total penalty of 2600 years in prison, it is formally well possible that Garavito will be released out of prison within the next 10 to 20 years, i.e. even before the maximum sentence of 40 years will be over.     

Aging and homicide rates

There are no published studies concerning the effect of mortality rate on the rate of homicide by habitually violent offenders. On the basis of nation-wide statistics in Finland, the frequency of homicide is 61% lower among 50-year old males than among 30-year old males. However, when the 4.9-fold mortality of habitually violent offenders is taken into account, the homicide rate is only 43% lower among 50-year old males than among 30-year old males. This is an important issue in forensic psychiatry and that requires further attention, as age is used a predictive factor when assessing the risk of violent behavior.     

Understanding sexual homicide

Abstract

Sexual homicide is a crime that receives much media attention, but one that still has a relatively small research base compared to other crimes. Although the idea of someone deriving sexual gratification from or during the killing of another is an anathema to many, forensic professionals need to understand sexual homicide and the range of circumstances and motivations that lie behind it. This paper explores this complex crime: it illustrates the various definitions and labels used; it suggests that sexual homicide is rare and is committed by individuals who, although having troubled histories, may not differ from other offenders. Then, by examining the various legal and behavioural circumstances in which sexual homicide can occur, this paper illustrates the diversity of sexual homicide and presents it in a more understandable manner. Finally, the paper suggests that knowledge of this diversity is required both for investigative and clinical purposes.     

Segregation and latino homicide victimization

A growing concern exists that an increase in Latino urban violence is the result of social and economic inequality. One structured form of inequality is segregation. Research indicates that many Latino communities have moderate to high levels of segregation. Prior criminological research has revealed that segregation is a strong predictor of black violence. The present study extends this line of research to the issue of Latino crime by examining the link between Latino segregation and Latino homicide victimization. Two measures of segregation are employed in the current research: residential segregation and social isolation. Using census data and mortality files, regression models indicate that while social isolation is a significant predictor of Latino homicide victimization, residential segregation is not significantly related to Latino rates of homicide victimization. An earlier version of this paper was presented at the annual meeting of the Southern Criminal Justice Association, Clearwater, FL, 2001.     

Spousal homicide and the subsequent staging of a sexual homicide at a distant location

The case of a 63-year-old man who killed his 52-year-old wife and then staged a sexual homicide at a distant location is reported. A review of all evidence, a forensic psychological interview, and psychological testing indicated that the murder was the result of a narcissistic rage reaction during which the subject beat his wife to death with a paint can, a clothing iron, and a rock. He then drove her body to a field 87.3 miles away, and positioned it in a manner that exposed her breasts and her underwear. He turned himself into the police two days later. There is no controlled empirical research on staging, although this single case supports the criminal investigative theory that staging exists, and is done to deliberately mislead homicide investigations (Douglas et al., 1992).     

Erotomania, triangulation, and homicide

A case of homicide by a 29-year-old male with erotomanic delusional disorder and various personality disorders is reported. Following a month of pursuit of a female stranger with whom he had briefly conversed in a local bar, he assaulted an automotive plant where she worked, delusionally believing that she was at imminent risk and needed to be rescued. One plant manager was killed and two police officers were wounded. The case illustrates the phenomenon of triangulation, where rage toward the rejecting object is displaced onto a third party, which is then perceived as impeding access to the victim and may be at risk for violent assault.     

Psychopathology and weapon choice: a study of 103 perpetrators of homicide or attempted homicide

Bone marrow (BM) analysis is of forensic interest in postmortem toxicological investigation in case of limited, unavailable or unusable blood samples. However, it remains difficult to determine whether a drug BM concentration is therapeutic or represents overdose, due to the lack of studies on this alternative matrix. Given the variations in BM composition in the body, sample location was suggested to be a relevant factor in assessing BM concentration. The aim of the present study was to compare postmortem caffeine concentrations in various BM sample locations and secondly to consider the correlation between BM and blood concentrations. Six BM samples (right and left side: proximal and medial femur and 5th rib) and a blood sample were collected from 21 forensic autopsies. Gas chromatography coupled to tandem mass spectrometry was performed. Blood caffeine concentrations ranged from 60 to 7591ng/mL. Femoral and rib BM concentrations ranged from 51 to 6171ng/g and 66 to 7280ng/g, respectively. Blood concentrations were always higher than BM concentrations. As a good correlation was demonstrated between blood and rib BM and between blood and the average of the four femoral BM concentrations, blood caffeine concentrations could be correctly extrapolated from BM concentrations. BM caffeine concentration was found to depend on sample location. Rib BM caffeine concentrations appeared to be systematically greater than averaged femur values and concentrations were much more variable between the 4 femur BM samples than between the 2 ribs. From a practical point of view, for caffeine analysis, rib BM appeared more relevant than femoral BM, which requires multisampling to overcome the concentration variability problem.     

Suicide and homicide away from home

In 1980, suicides in the United States that occurred outside the victim's home state were most often committed by young single or divorced men. Hanging was the most common means of suicide. Homicide victims who were killed outside their home state were usually male, black, divorced, young, and killed by unusual (other) means or by law-enforcement officers.     

Work-place homicide by bow and arrow

Arrow wounds represent an unusual class of wounds rarely seen by most forensic pathologists. In this paper we present a case of homicide by bow and arrow and the characteristics of such injuries. The essential characteristics of the lesions obtained from conically-tapered field points and from hunting broadhead tips are described and discussed in relation to injuries caused by firearm bullets. In the present case, three arrows struck the victim, and the order in which the injuries were sustained are analyzed. We also discuss the possibilities of localizing the shooter relative to the victim by analysis of the trajectories.     

Execution publicity and homicide in Georgia

A requisite of both general deterrence and brutalization theories of capital punishment is citizen awareness of executions. Yet this dimension is frequently left unmeasured in the literature on executions and homicide. Further, the current literature utilizes nationwide data, which introduces aggregation bias. The present study deals with these issues, testing the brutalization theory with data on publicized executions in the state of Georgia. Monthly data are used in order to capture any short-term brutalization effect. The results of a multivariate time-series model find that a publicized execution is associated with an increase of 2.6 homicides, or 6.8%, in the month of the publicized execution. Publicized executions were associated with an increase of 55 homicides during the time period analyzed. The results provide no support for the deterrence, victim mobilization, and normative validation perspectives on homicide. The findings can be interpreted, however, from several different varieties of brutalization theory.