Monday, March 26, 2012

How Best To Count Your Crimes

This is a bit of an odd one. There was a story this weekend out of the Galveston Daily News that had this interesting tidbit buried in it:

Some of the figures might not reflect an exact number of crimes reported, as Galveston police changed the way it reports crimes to the state and FBI. The department used an incident-based reporting system for half of 2011, police Lt. Bryon Frankland said.

When compared to the FBI’s Uniform Crime Report, however, the numbers would be different because the FBI excludes multiple offense data derived from the same instance of a crime. For example, if a person is killed during a burglary, the FBI lists only the homicide in its Uniform Crime Report.

I think there is a couple of things worth noting, one I'm a little puzzled why Galveston Police would change from reporting crimes using the National Incident Based Reporting System (NIBRS) to Uniform Crime Reports (UCR) mid year. I'm not puzzled about why an agency would switch from NIBRS to UCR. NIBRS is generally seen as more complicated than the much older UCR system.

Of course, the UCR system which was designed in 1929 has it's share of faults. NIBRS which was designed in the late 1980's is probably a better system but given it's additional complication, there is no compelling reason for law enforcement agencies to adopt it.

I'm interested in why an agency would make the switch from NIBRS to UCR or what reasons are there to switch from UCR to NIBRS?


  1. In Washington State the change is mandated by the Association of Sheriffs & Police Chiefs. Participation in crime reporting if voluntary but as of 01/01/2012 reports must use NIBRS.

  2. Seventy, Thanks for the comment. While I think there is probably a better way to count crime than UCR, I'd hate to be forced to move to NIBRS. It would add a significant complication to our workflow and I am sure that the state would not help us defray the additional costs.

  3. I think perhaps the "multiple-crime-counting" issue has been a bit oversold. First off, a lot of people have the wrong impression of it. NIBRS still recognizes the concept of "lesser included crimes"; for instance, burglaries do not get associated larcenies, nor robberies associated assaults (I've heard people claim both). Situations in which someone is murdered during a burglary or raped during a robbery are rare enough that I doubt it would greatly influence an agency's statistics.

    I recently did an analysis for a mid-sized city that was concerned about the issue. Out of 1,956 incidents that had at least one Part 1 crime in 2011, only 16 had more than one. In 12 of those cases, the lesser crime was a type of larceny. I looked back until 1998 with their data, and the highest number of multiple-Part-1 incidents they had in any given year was 20. In no year did it exceed 1% of their total Part 1 crimes. My conclusion was that the increase caused by switching from UCR to IBR would be negligible. This is not to mention that they build reports to continue to report under the old standards while still submitting IBR data to the state and FBI.

    The bigger issue with IBR is that by reporting incident-level data, and multiple data fields per incident, the system requires agencies to exert a much greater degree of quality control over the data. Under UCR, an agency could classify a crime any way it wanted and just report the summary totals. Only an audit would reveal the mis-coding. Under IBR, the validation rules won't allow an agency to report a robbery without also reporting what was stolen (attempts excepted), or an aggravated assault without a weapon or serious injury. Generally, such agencies need someone to review each submitted report and ensure that it meets the incident type definitions and the business rules for coding. If an agency just doesn't have the resources to invest in that process, I could see them switching back, but I guarantee their data quality will take a hit.

  4. Chris, At my agency we already spend a great deal of time on quality control just for UCR. I'd hate to have to spend even more time on it. I think that if you give agencies a choice, they are going to take the easiest path. It would probably an effort similar to the one in WA to get everyone in Texas on NIBRS. Few agencies are likely to make the move all on their own.


I reserve the right to remove defamatory, libelous, inappropriate or otherwise stupid comments. If you are a spammer or are link baiting in the comments, a pox be upon you. The same goes for people trying to sell stuff. Your comment will be deleted without mercy.