Remember the old doggerel about if a tree fell in the forest and nobody heard it, did it really fall? I’m running into the same kind of problem in trying to understand the data on crime.  There are two agencies that publish crime data: the FBI (Uniform Crime Reports) and the BJS (National Crime Victimization Survey.) With one exception, all of this information comes from statements by crime victims who may or may not choose to report the crime. The one exception is homicide because it’s pretty tough to hide a dead body plus, given the severity of the crime, the moment we even think it has taken place, everyone gets into the act.  Otherwise, there isn’t a single category of serious (or non-serious) crime whose occurrence can be counted or even estimated without the cooperation of the victims themselves.

I have been trying to figure out how many crimes really take place for two reasons. First, the question has become a big political football in the ongoing debate about guns.  The NRA and its allies claim that the drop in violent crime over the last twenty years demonstrates both the futility of more gun laws and the efficacy of concealed-carry permits as a further defense against crime.  The gun control crowd, on the other hand, points to the fact that although the overall rate of serious crime has declined, the homicide rate due to the proliferation of guns, is still much higher than we would like.

The second reason that I have been trying to figure this out lies in the disparity between crime data generated by the FBI as opposed to crime victim data produced by the BJS.  The gap between those two reports has narrowed considerably over the last number of years, but it is still significant enough to make me wonder whether the numbers can be trusted at all.  As a starter, let’s compare crime data for 2012, the most recent year for crime data published by both agencies.  According to the FBI, there were 1,214,462 homicides, forcible rapes, robberies and aggravated assaults committed that year. According to the BJS, there were 2,084700 serious criminal victimizations that same year, and this number does not include the nearly 15,000 homicides reported by the FBI.  Now according to the BJS, virtually all the victmizations covered by their survey are reported to the police, but I since the data for this assertion is presented in terms of rates per 1,000 rather than raw numbers, I can’t really figure out why such a discrepancy between between the two reports exists.

And the discrepancy becomes much greater if we go back to the period when, according to both agencies, there was a lot more crime.  Let’s look at the data for 1996, which is considered the high-water mark for crime levels over the last two decades.  According to the FBI, there were 1,688,540 serious crimes reported in 1996, the number of 1996 victimizations, according to BJS, was 3,371,445 (adding the murders counted by the FBI.) In that year the difference between BJS and FBI numbers was 2:1, again, a discrepancy which neither agency seems able to explain.

But what this might explain are all the public polls which indicate that most people believe that violent crime in on the rise, even when the official numbers keep show that it is going down.  In a survey published last year during the debate over a new gun control law,  Pew found that a majority of Americans (56%) believed that crime was at higher levels than during the 1990’s, and only 12% thought it had gone down.

The difference between the data from the FBI and the BJS can’t just be dismissed as stemming from different definitions of crime or different methods of  data collection or different something else. You can, in fact, read a very detailed statement about the difference between the two sets of data published by the Department of Justice (which oversees both agencies) but it doesn’t offer even the slightest acknowledgement that the disparity in numbers published by the two agencies calls into question the accuracy of either one.