Using Behavioral Patterns to Link Serial Rape Offenses:A Multidimensional Approach

The purpose of this project is to help improve the law enforcement investigative practices for serial rape by identifying empirically based and methodologically sound ways of linking crime series. An effective linking process (e.g. the timely recognition that multiple crimes are part of a series) could expedite rape investigations and assure the successful capture of the suspect in a timely fashion.

The current data set consists of 29 series with a minimum of three and a maximum of 15 crimes per series. All crimes were committed by male offenders acting alone, who together are responsible for a total of 192 victims (i.e. distinct assault incidents). Variables that will be used in this study were coded using the Homicide Profiling Index, Revised for Use with Rape (HPI-R, Salfati, 2010), an instrument specifically devised for use with police files.

The study aims to: 1) identify the behavioral trajectories within the three key behavioral dimensions of rape offenses identified in the literature; namely violence, sexual activity, and control; 2) determine whether these patterns or trajectories can be used for correctly linking together crimes that belong to the same series. Aim #1 will be accomplished using a multidimensional scaling technique called Partial Order Scalogram Analysis with Coordinates (POSAC) that allows for both quantitative and qualitative differentiation. Once the individual dimension trajectories are identified, configural frequency analysis (CFA) will be used to determine the most and least common cross-dimensional trajectories. This information will then be used to accomplish Aim #2, using a multinomial regression analysis to determine whether the identified trajectories can be used to successfully link crimes. ca/ncf