Page:FBI Law Enforcement Bulletin 54 (8).pdf/15

 Chapter 3 Classifting Sexual Homicide Crime Scenes Interrater Reliability

The unsolved homicide presents a major challenge to law enforcement officers. These unsolved cases, which often include a sex-related compo­ nent, usually have no apparent motive. The victim has been sexually abused, and the nature of the killing indicates behavior patterns that reflect sexual deviation, specific character traits, and perhaps even psychopathology. Also referred to as lust murders,$1$ these murders often include severe beating and multiple stabbing of the victim, body mutilation (such as removal of sexual organs), and sexualized positioning of the body after death.

The FBI's Behavioral Science Unit (BSU) has been involved since 1972 in assisting city, county, and State law enforcement agencies in their investigations of unsolved murders by preparing profiles of the unidentified offenders, after extensive examination of the crime scene data, victim characteristics, and autopsy reports. This profile may include the perpetrator's age, race, sex, socioeconomic and marital status, intellectual and educational level, occupation, lifestyle characteristics, arrest history, location of residence in relation to the scene, and certain character traits.

The Agents responsible for preparing the offender profiles have found it useful to classify the type of crime and the organizational structure of the crime scene. The crime is classified as sex-related, nonsexual, or unknown. Evidence of a sexual component anywhere within the crime scene justifies the sex-related classification. The organizational structure of the crime scene is determined by evidence of the amount of planning and premeditation by the offender, as well as of the offender's control over the victim. For example, a weapon taken to a crime scene and carried away suggests planning, as contrasted with a weapon used and left at the crime scene, suggesting opportunity and spontaneity.

In sex-related crimes, the structure of the crime scene provides in­ sight into the offender's patterns of behavior. For example, a well-organized crime scene indicates an of­ fender with a conscious plan of action after the murder to avoid detection and apprehension.

Currently, the BSU is systematically studying their profiling procedures through scientific and statistical analyses. Because of the importance of correctly classifying the crime and the crime scene, we needed to establish the reproducibility of these classifications. This article reports our investigation of the Agents' ability to reproduce independently each other's classifications. This ability to replicate decisions is called interrater reliability.

Study Design Six BSU Special Agents with varying levels of experience in profiling participated in the reliability investigation. Data from 64 murder scenes, covering a variety of circumstances both sexual and nonsexual, were selected for the study. For each crime scene selected, one of the participating Agents was thoroughly familiar with the case. This Agent presented a short description of the crime scene and showed crime scene photos. The presentation was restricted solely to information immediately available at the crime scene; no information from laboratory tests or later investigation was divulged. This restriction allowed the other Agents to focus on immediate data. Other details of the investigation, if discussed by the presenter, might have influenced the Agents in forming their

August 1985 / 13