Monday, November 19, 2007

New Item from the Department of Bad Statistics

The FBI's annual hate crimes report came out today, reporting a jump of about 8% in reported hate crimes between 2005 and 2006. Predictably, Al Sharpton has jumped on this, saying "The FBI report confirms what we have been saying for many months about the severe increase in hate crimes." Presumably, he was referring primarily to a perceived increase in racial hate crimes, since he included in his remarks a call for AG Mukasey to meet the Congressional Black Caucus, and since Sharpton has hardly been the most sensitive person in the world when it comes to stoking religious hate crimes (see, e.g., Freddie's Fashion Mart, Crown Heights, and his recent remarks about Mormons).

Anyways, there's a huge problem with the statistics: they're devoid of meaning. First of all, there is of course the issue that the term "hate crime" is almost entirely subjective, since it requires an exploration of the perpetrator's motives.

Second of all, there is the fact that the statistics may therefore just reflect an increased sensitivity to intolerance of various stripes- especially against gays (who, as Andrew Sullivan has regularly pointed out, are becoming increasingly accepted in American culture, with some obvious exceptions). In other words, increased acceptance of certain groups inherently makes it more likely that someone will label a crime against them as a "hate crime." This conclusion is supported by the fact that the two types of "hate crime" that increased as a percentage of overall "hate crimes" were sexual orientation and religious based hate crimes (which it is worth pointing out are a particular point of emphasis in the Christianist-run DOJ).

Third, the total number of reported hate crimes in the report is only 7722 from 12,620 reporting agencies- only a little more than 1/2 of an incident per reporting agency, and only one incident for ever 38,000 or so Americans.

But most important of all is this little nugget buried at the very end of the article: "Lack of full participation by the more than 17,000 police agencies around the nation somewhat undermines year-to-year comparisons.For instance, in 2004, 12,711 agencies reported 7,649 incidents. In 2005, only 12,417 agencies reported and incidents dropped 6 percent to 7,163. But in 2006, agencies reporting rose to 12,620 and incidents climbed 7.8 percent to 7,722."

In other words, once you account for the number of agencies reporting, that nearly 8% increase from 2005-2006 drops to just a little over 3%. It also represents a statistically insignificant increase of just about 1% over 2004.

Another little factoid that Rev. Al and CNN missed: the rate of race-based hate crimes actually stayed about identical from 2005-2006. The article indicates that the percentage of "racial bias" crimes fell "slightly" from 54.7% to 51.8%. This "slight" drop means that the reported number of race-based "hate crimes" increased only from 3918 to 4000. Once you factor for number of agencies reporting, that means that the number of race-based incidents per reporting agency stayed the same from 2005-2006, at just about .32 incidents.

A few other things: the report cannot account for the possibility that law enforcement is increasingly likely to classify an event as a hate crime due to the proliferation of hate crimes laws in recent years, which can add time onto a suspect's potential sentence, and give prosecutors an additional tool to force a plea bargain. Also, the proliferation of hate crime legislation may mean that incidents that previously were not crimes at all (and thus unreportable) are now classified as crimes (and thus are reportable).

The report also indicates that only 58% of perpetrators were white (with another 13 percent of undetermined race). Given that 51.8% of incidents were classified as "race-based", this means that either: 1. Whites perpetrated an extremely small percentage of anti-gay and anti-religious "hate crime" (extremely unlikely), or 2. A sizable percentage of race-based "hate crimes" were perpetrated by nonwhites, or 3. Prosecutors are disproportionately unwilling to charge whites with race-based hate crimes. If the answer is 2, then we are completely overemphasizing white on black hate crime in this country, and Al Sharpton should give the protest march thing a bit of a rest. If the answer is 3, then we have no freaking clue how much of a problem hate crimes are in this country and the statistics are meaningless one way or another.

Finally, the CNN article points out that the Jena case is not included in the statistics. While this is unfortunate, it is statistically irrelevant- we're talking about a handful of crimes (whatever their relative importance) out of thousands. That's not going to change the statistics in any meaningful way.

In conclusion: the media has not surprisingly jumped all over these statistics. Unfortunately, and as with most statistics cited by the media, they are mostly devoid of meaning. To the extent they have any meaning at all, it is a very different meaning from that portrayed in the MSM.

Just some food for thought...