Wednesday, January 22, 2003

I agree with the Professor about consistency. The Paulos analysis of TIA which makes it unpalatable also makes gun control unpalatable. But let's make sure the comparison is indeed consistent.

After assuming that 3% of population will perpetrate a violent crime and that in 8% of the crimes a gun will be used, we get roughly .2% of population using a gun in a violent crime. Assuming gun owners are no more criminal than the population as large and that about 25% of total population are gun owners, we could expect roughly 150,000 perpetrating gun owners in a population of 75 million gun owners.

So what is the rationale of gun control? If reduction of gun ownership is the goal we would have to drop 500 law-abiding gun owners to stop 1 criminal gun owner. But supposedly the intent of gun control is to increase that ratio. In theory the debate over gun control should center around whether the degree of intrusion and inconvenience to 500+ people is worth it to apprehend just 1 criminal. I suspect that there are more cost-effective ways to accomplish the same goal.

On the TIA thing it takes processing over 3000 people to find 1 terrorist.

I wonder how many other issues could stand up under a similar treatment.

There is a missing statistic in the gun control scenario. With TIA we assumed a 99% effectiveness. This gave us an idea of how many would be incorrectly profiled as terrorists. In gun control we don't have a stat that reflects how many gun owners would profile as potential offenders, either correctly or incorrectly. The proper ratio we need to compare to the TIA ratio is the number of incorrectly profiled gun owners to the number of actual offenders. If for the sake of argument we assumed the same effectiveness as in the TIA example the ratio becomes more like 5 to 1 instead of 500 to 1.

It should also be noted that we assumed 1000 terrorists in the TIA example. There is a lot of leverage in that number. If there were only 500 terrorists the ratio shoots up to 6000 incorrect profiles to 1 terrorist.

This make gun owner profiling look like it's worth doing. How about it, Professor?

No comments: