Perhaps one of the reasons police keep beating people senseless across America is they don’t know what a serious assault is.
The Los Angeles Police Department (LAPD) has misclassified the seriousness of around 14,000 crimes over eight years, beginning in 2005, according to an investigation by the Los Angeles Times. There were really 7% more violent crimes and 16% more serious assaults than reported.
The report builds on a Times investigation in August 2014, which concluded that 2,000 serious crimes by FBI standards—these numbers go to Washington—were misclassified as minor between October 2012 and September 2013. Seventy percent were violent and the rest were property crimes. After running a sampling of its data past a team of experts, the paper reduced the number to 1,200 and showed the results to LAPD officials, who concurred and said they would fix the reporting.
The department did put a number of new procedures in place, but the Times reported earlier in the week that classification errors for at least aggravated assaults continued through 2014. An internal LAPD audit found there around 23% more than originally reported by the department.
LAPD uses a data-driven approach to policing that bases decisions like budgeting and officer deployment on state-of-the-art analysis of crime statistics. The system, CompStat, was championed by William Bratton in 1994 when he was police commissioner in New York City and he brought the system to L.A. when he led LAPD between 2002 and 2009. Bratton returned to New York, where he leads NYPD again.
Like any computer-based system, garbage in means garbage out, and apparently the city has made a lot of important decisions based on misinformation. While the new, improved assault statistics are much higher than the originals, they do not reverse what has been reported as a steady downward trend. They do, however, straighten out the slope of decline.
No doubt, some of the errors were innocent mistakes. But the Times littered its report last August with cases that were clearly egregious errors or willful deceit. A man stabbed his girlfriend with a scissors; another beat and stabbed his girlfriend with a screwdriver; another man beat a neighbor senseless with a metal bar. All were deemed minor offenses in the reports, although the prosecutor and courts tended to think otherwise.
The Times found a lot of the mistakes by searching through reports for words like “knife” and “stab.” If they were innocent errors, one would expect to find a similar pattern in less serious crimes. But that was not the case.
CompStat is widely used and has been credited by many for helping drive down major crime by identifying criminal patterns and facilitating law enforcement responses. John A. Eterno and Eli B. Silverman, authors of “The Crime Numbers Game: Management by Manipulation,” documented in their book about NYPD how CompStat was used to evaluate the performance of officers and perhaps drive down crime stats, but not necessarily crime.
“Once you have one ‘CompStat' meeting where they're screaming and yelling at you about your crime numbers, you get the hint and then you do what you can to make sure those numbers are looking the way they want them to.”
The good news about LAPD’s bogus crime stats is that they were consistent, according to the Times. They made roughly the same percentage of errors from one year to the next, so the relative trajectory of crime’s decline was accurate. LAPD just had the number of serious crimes underestimated.