Paulsen demonstrates that by improving the commuter/marauder prediction early in the process and by conducting more targeted research in the field, geographic profiling technology can be more effective and results can be more accurate. Geographic profiling has been used in crime analysis since the early 1990s. It is an investigative methodology that relies on location of connected crimes series as a means of determining the next probable location known as an offender anchor point. According to Dr. Paulsen, 20 to 60 percent of crime series are committed by commuters, and geographic profiling is highly inaccurate in predicting these series.
Therefore, it is important to analyze variables and methods used to classify commuters and marauders prior to geographic profiling. This study analyzed 106 crime series gathered from 25 U. S. jurisdictions. This study differed from traditional methods in the use of variables applied to classify offenders as commuter or marauder. Traditional models use residence as primary variable to determine classification. Paulsen modified this method to look at any of 14 known variables which might classify as marauders where traditional models may classify as commuters and performing logistic regression with SPSS software.
Since profiling is inaccurate in predicting commuter series, Paulsens results improved the accuracy of prediction rate from 60% to 81% by changing the classification process. The findings are significant in that they are at odds with traditional theories of offender patterns and characteristics. Traditional theories state that marauders commit offenses closer to home. However, Paulsens research shows that commuters operate in smaller geographic areas with more locations while marauders operate in larger more dispersed areas. Three important implications arise from this study.
Further research is needed to develop a method for determining commuter/marauder classification that can be used in multiple geographic profiling systems. Research is needed to determine if methods and results are accurate in various jurisdictions with differing demographics and crime characteristics. Alternative profiling and forecasting methods must be weighed against this and other existing research. Previous studies were limited in number, only two, and in scope, focusing on software platform features and case studies and covering limited jurisdictions.
This study more extensive in size and scope. The primary limitation of this study is that the entire sample consists of previously solved crimes. This underscores the need to conduct similar research on unsolved crimes to determine the accuracy of this model when all variables are not known. Additionally the case selection process for this study is limited by the police departments rationale for selecting cases to send for the study which is not addressed. Geographic profiling is a widely used crime analysis methodology which can and should be more effective.
By analyzing existing results and modifying components of the profiling process crime prediction rates can be significantly improved. Improving results of geographic profiling will depend on continued research of adequate size and scope that focuses on alternative classification methods, a range of geographic locations, and solved and unsolved crimes. Reference Paulsen, D. (2007, September). Improving Geographic Profiling through Commuter/Marauder Prediction. Police Practice & Research, 8(4), 347-357. Retrieved March 11, 2009, from Academic Search Premier database.