When is an 8% rise not really an 8% rise?

Before people get too worked up about the news that hate crimes are on the rise you should be aware of this crucial bit of information:
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.

Now why this, the most important piece of information in the whole story, is buried at the end of a rehashing of the Jena 6 and various noose incidents is beyond me. These four short sentences tell us everything we really need to know about the so called 8% rise in hate crimes. With a different cohort of agencies reporting each year there's no way to tell if there is any overall change in the rate of hate crimes.

You can't determine a rate of change in anything if you change the nature of the entity that you're observing. Claiming that there was an 8% rise in hate crimes based on these numbers is like measuring the height of two different trees and then claiming that the difference between the two shows that the shorter tree shrunk.

Now if someone chose to look at the numbers from the police agencies that:

  • consistently report hate crimes from year to year

  • and

  • all consistently use the same criteria to determine what is a hate

  • then you could determine a rate of change for hate crimes in those particular agencies. If the number of agencies that met the above criteria was large enough and if those agencies were widely dispersed across the country then one might be able to extrapolate some sort of national trend in the rise or fall of hate crime occurrences. To do anything else is a sloppy waste of time and money leading to sensationalist but totally false headlines like "Hate Crimes Rose 8 Percent in 2006"

    Somebody needs to take some remedial lessons in statistical analysis and application of the scientific method.

    Comments

    1. Further, different states have different definitions of hate crimes, and each jurisdiction has its own interpretation of its state's laws.

      Plus, if anyone does want to use these numbers for a debate, check this:

      59% of hate crimes were by whites (80% of the total population), but 21% were by blacks (12% of the population).

      So, umm, hmmm....

      ReplyDelete
    2. The subject of hate crimes have been politicized and taken off track as a whole. Certain groups want attacks upon them labeled as "hate crimes" while staying silent when a member of their own group attack another person who is across racial lines.

      http://withintheblackcommunity.blogspot.com/2007/07/impact-of-hate-crimes-prosecutions-on.html

      As a Black person I can't help but note the past in which so many laws that were originally advocated for by people interested in the wellbeing and protection of Black people have later become the means by which Black people have been prosecuted. It is my belief that hate crimes will be the same. The increasing conflict between Blacks and Hispanics will one day lead to these two groups being the most frequent recipients of Hate Crimes prosecutions.

      ReplyDelete
    3. I've always been dubious about the separation of certain instances of breaking the law into "hate crimes" categories. But that's a separate issue from the problems with the generation of this 8% number.

      ReplyDelete

    Post a Comment

    Popular Posts

    What's in a flower?

    Treating autism as traumatic brain injury

    Battlestar Galactica needs more colour?