NIJ does not exercise control over external Web sites. Read our Exit Notice.
Geographic profiling is a technique that can help identify the likely area where a serial offender resides, or other place
(e.g. work, girlfriend's place) that serves as an anchor point or base of operations.
Several major software programs are being employed by police agencies to perform geographic profiling tasks, including Rigel®, CrimeStat, and Dragnet.
Though there have been anecdotal successes with geographic profiling, there have also been several instances where geographic
profiling has either been wrong on predicting where the offender lives/works or has been inappropriate as a model. Thus far,
none of the geographic profiling software packages have been subject to rigorous, independent or comparative tests to evaluate
their accuracy, reliability, validity, utility, or appropriateness for various situations.
In August 2004, MAPS contracted with Abt Associates to convene a geographic profiling roundtable panel, bringing together
ten independent experts with backgrounds in geography, criminology, crime analysis, software development, and spatial analysis,
to derive a comprehensive evaluation methodology of geographic profiling software. The main evaluation criteria is how accurately
does the tool predict the offender’s 'base of operations' - their home, job, or other frequented location? The evaluation
must be conducted using crime data from jurisdictions with varied geographic characteristics, offender types, and offender
travel patterns. Other evaluation factors include software features and usability.
The final report A Methodology for Evaluating Geographic Profiling Software (pdf, 259 pages), produced by Abt Associates, details the geographic profiling evaluation methodology.
Response to the methodology
CrimeMap Listserv discussion
Since the report was released in January 2005, debate has ensued among the geographic profiling software developers, researchers,
and practitioners, as to the validity of this methodology, the various proposed output accuracy metrics, as well as the strengths
and weaknesses of the software packages, the limitations of geographic profiling, and how to advance geographic profiling
through further research.
Kim Rossmo's Response to NIJ's Methodology for Evaluating Geographic Profiling Software
Kim Rossmo prepared a response to the National Institute of Justice’s A Methodology for Evaluating Geographic Profiling Software:
Final Report (Rich & Shively, 2004).
His response, in summary:
The report contains certain errors, the most critical of which involves suggested performance measurements. Output accuracy
is the single most important criterion for evaluating geographic profiling software. The report discusses various performance
measures; unfortunately only one of these (hit score percentage/search cost) accurately captures how police investigations
actually use geographic profiling. This response addresses the various problems associated with the other measures. Geographic
profiling evaluation methodologies must respect the limitations and assumptions underlying geographic profiling, and accurately
measure the actual function of a geographic profile. To evaluate geographic profiling properly requires analysing only those
cases and crimes appropriate for the technique, and measuring performance by mathematically sound methods, in order to meet
NIJ’s standard of a 'fair and rigorous methodology for evaluating geographic profiling software'. -- Kim Rossmo (4/20/05)
Kim Rossmo's full response is available for download from the Texas State University's Center for Geospatial Intelligence and Investigation website.
Response to Kim Rossmo’s Critique of the NIJ Methodology, from Ned Levine
This discussion of geographic profiling methodology, stimulated by Kim Rossmo's critique of the NIJ evaluation, prompted Ned
Levine to write some comments. He has posted them on his web page. (Posted: 5/11/05)
Date Modified: December 15, 2009