Saturday, March 23, 2013

Expensive Study, Trivial Results

Yep, there's another politically biased study out of Harvard, this time by an Eric Fleegler, about guns.

Here's a chearleader's view from Boston's NPR station, which doesn't seem to be able to understand the statistics involved.

Here's the scant praise that comes for the study from Garen Wintemute, an emergency medicine physician and director of the Violence Prevention Research Program at the University of California, Davis:

"In the end, Fleegler and colleagues provide no firm guidance and leave us with more questions than answers. Do the laws work, and if so, which ones? Should policymakers enact the entire package, or just some of the measures?" he said.
Professor John Lott had a more pungent analysis here

But the report is based on embarrassingly bad statistics that are rigged to get the result the authors wanted.
How would that work? I mean, it's a scientific study, how could it possibly be wrong?

In addition, we calculated household firearm ownership rates per state using the firearm suicide/total suicide ratio, which is the proportion of all suicides in a state caused by firearms.
Okay, so we're first going to figure out what percentage of gun ownership is by using the suicide rate (with guns) as a proxy.

Then we're going to calculate how much gun control laws reduce crime by comparing the gun violence rates (including suicides) between the various states.

And the result, of course, is that more gun ownership means more violence, because the numbers above can't come out any other way. If people are going to commit suicide anyway, easy access to buying guns means more of the suicides are accomplished with guns. That increases both the "household firearm rate" and the "overall firearm violence rate".


The study authors admitted that the difference only became significant if you compared the bottom quartile to the top quartile.

The study authors conveniently omitted Washington DC, which has extremely onerous gun laws and a sky-high gun violence rate.

The study authors conveniently did not omit Louisiana, which has lax gun laws and a sky-high gun violence rate.

(According to Power Line, dropping out louisiana from the study results in the remaining bottom 25% in gun laws being 12% safer than the top 25% in gun laws.) The study made no attempt to figure out how much violence there may have been that gun ownership prevented.

Cheat, much?

No comments: