6/28/2006

Nemerov: "AP Blames NRA for Violent Crime"

Howard Nemerov over at ChronWatch has an interesting article on media bias: AP Blames NRA for Violent Crime. Howard is right to point to the bias in who reporters rely on for their experts. I think that some of the points go too far, such as worrying about the fact that the initial crime numbers don't cover the entire country or that there are frequently small revisions in these initial numbers. I also wouldn't push the political bias that hard.

There is also the issue of timing. I thought that the federal funding was cut well before this increase occurred. In addition, the local communities bear the costs and benefits of crime and they are perfectly capable of deciding how much of their own money to spend on law enforcement. It is not really clear why you have localities send their tax money to the federal government only to have it returned to them with various strings attached.

2 Comments:

Anonymous Anonymous said...

Several important issues were not dealt with in the Blame article.

Are the crime rate increases statistically significant?

In the computation of statistical significance, what is the right probability to compute: The probability of one year with such an increase or the probability that at least one year in "n" will have such an increase? Your answer depends on what motivated the writer. My bet is that the writer was motivated by examining the past and finding a year that looked relatively bad. Thus, I think the latter question is the right one. Any reasonable computation of that probability will surprisingly large. No likely news there.

If you wait long enough, there will be a new contender for both best and worst outcome. This, by itself, means nothing.

Robert Ferguson, Ph.D.
NRA Certified Instructor

6/28/2006 4:34 PM  
Blogger John Lott said...

It is indeed hard to get a statistically significant result with only one year. But I don't think that they are even thinking about statistical significance.

6/29/2006 3:59 AM  

Post a Comment

<< Home