The Need

A Southwestern city identified the need to determine if officers were targeting minorities in two public mall locations.  Allegations concerning the targeting of Hispanic males had surfaced, indicating concern in the community that inappropriate stops and detentions were being made.  The allegations came primarily from anecdotal information from the community, however, some preliminary statistics showed that a disproportionate number of Hispanic males were being stopped compared to the number of Hispanics living in the city.  The City needed a third party to determine whether officers patrolling those areas were indeed targeting Hispanic males, or whether the initial statistical disparity could be explained.  The need was critical, because at the same time that some community interests were charging racial profiling, mall security and other segments of the community were asking the police to reduce mall crime.

The Solution

The first key was to educate stakeholders on the fallacy of the simplistic statistics used, often referred to as a bi-variate analysis.  The appropriate comparison group is not the census data of the city, because the entire city does not traffic the mall at the same times and at the same frequency.  The appropriate group is the actual people who frequent the mall.  We explained that the best method for identifying this group is a direct measurement of the racial and ethnic breakdown of the people at the mall.

The second step was to break down the mall traffic by time of day and specific mall location.  That is, not all people traffic the same parts of the mall at the same frequency and at the same times.  This means that a much better, more accurate comparison has to do with racial and ethnic populations at different times, and at different locations within the mall.  It is intuitive that different locations within the mall would show that the racial and ethnic composition of mall traffic is not the same at all times and at all locations.

The next step was to review crime statistics and officer deployment strategies.  We found that officers were deployed in relation to calls for service and based upon crime analysis, which is a standard industry practice.  This means that officers patrolled locations of the mall at specific times, and in locations where more crime takes place.  Officers also responded to calls for service, in which specific descriptions of subjects were often given.  These descriptions resulted in Be On the Look Out (BOLO) situations in which agency policy dictates that race can be used as one factor in identifying criminal behavior.

The next step was to review behaviors that could trigger an officer’s attention.  We refer to this as the objective assessment of criminal activity.  This activity involves identifying and defining the characteristics, behaviors and key indicators that could reasonably predict or indicate violating behavior.  We established these behaviors through identifying what the officers look for, and then by benchmarking these activities against other agencies and industry standards.

The final step was to incorporate these factors into a method to measure the baseline comparison group.  Through the development of the right design and the right analysis, we developed the true baseline comparison group, and used this baseline to compare against the agency’s stop data.

Outcomes

As we suspected, we found that the simple city census drastically underestimated the ethnic composition of mall traffic.  That is there were a lot more Hispanic males frequenting enforcement areas than the census data alone predicted.  We also found that the stop percentages were much higher for Hispanics in areas of the malls that Hispanic males frequented.

Upon further review with the agency, we discovered that the calls for service at one of these malls identified young Hispanic males as the suspects approximately 58% of time, a proportion of Hispanic suspects that was considerably higher than either census data or our benchmarking would have suggested.  It was, however, roughly the same proportion of Hispanics who were stopped in this mall. 

Through the careful design and implementation of the right study, and through analysis of community requests and agency policy and procedure, we were able to develop an accurate assessment of disparity, and the rationale behind enforcement activity targeted at reducing crime.  The results of this analysis were to educate the agency and the community so that discussion moved from whether targeting was occurring, to whether the community was willing to live with the agency responses to their concerns about crime.