Early warning

   Research conducted by PARC and Bobb shows that early identification systems and subsequent targeted intervention reduce police misconduct. PARC research also has shown that as a result of having an early warning and tracking system, LASD has substantially reduced its exposure to litigation, and when the department does face litigation, the payout is reduced.

   Bobb applauded the LASD and its Risk Management Bureau for its "recent and relatively stable" decreases in lawsuits brought against the department. In its 25th semiannual report in 2008, PARC looked at the last six full years of litigation against the LASD from 2001-2002 to 2006-2007. (The county's fiscal year runs from July 1 to June 30.) PARC found the number of new lawsuits filed against the department was down. In 2001-2002, there were 282 new lawsuits. In 2006-2007, there were 208 new lawsuits. More importantly, Bobb says looking at fiscal year 2004-2005 through 2006-2007, the department averaged 231 new lawsuits per year, which is nearly an 18 percent drop.

   A good deal of an agency's success in managing risk may be said to reside in the number of lawsuits not filed, Bobb says. So the drop in the number of new lawsuits, he says, suggests the LASD's risk management strategies are making a difference.

   Force-related litigation can generate high costs, large settlements and negative publicity. Here, too, PARC found the number of new force-related lawsuits and the number of closed force-related lawsuits resulting in payout were trending down. There were 78 new force-related lawsuits in 2001-2002 and 66 in 2006-2007.

   While the numbers have moved up and down, a comparison of three-year averages reveals that generally the number of new force-related lawsuits has trended down. In the most recent three-year period, the department saw an average of 62 new force-related lawsuits per year. The number of force cases requiring a payout by the county have gone down. And while the overall numbers of cases requiring payment have fluctuated, the LASD lost or settled an average of 23 cases per year for the most recent three-year period of the study, compared to about 30 per year for the similar preceding period.

   More importantly, Bobb says the percentage of force cases that the county closes, whether by dismissal, verdict, or settlement, that require at least some payout to the plaintiff, appears to be trending down. The average percentage of cases that required payment, due to settlement or verdict against the department, was about 35 percent in the three-year period from fiscal year 2004-2005 through 2006-2007, compared to the nearly 44 percent of cases that required payout in the preceding three-year period from 2001-2002 through 2003-2004.

   In looking at the effectiveness of the PPI, PARC used multiple regression analysis, which does not allow for predictions for specific officers. In bold print, the study says, "We cannot predict that Officer Jones will become involved in a shooting because he has five allegations of excessive force. We can say that five allegations of excessive force for officers in general are related to officers being involved in an additional shooting."

   PARC's 2009 research looked at 561 randomly selected, full-time officers with at least three years' worth of data in the PPI database.

Why it works

   The 27th semiannual report explains why the PPI works: Substandard behavior does not usually exist in isolation or occur randomly. It is related to other behavior that the PPI captures systematically. In other words, a good badge isn't going to tarnish overnight, but it's going to show signs of not being taken care of. The study shows an officer that becomes involved in an incident that's especially harmful to the department or community will have a history of problematic behavior.

   The PPI not only reveals these relationships, but also the magnitude of these relationships. It identifies behaviors that are typically associated with other potentially problematic behavior that the LASD wants to minimize.

   Specifically: