Interesting (not that we did not know this already)

[avs11054]

I am taking a research class for my masters degree. I have been wanting to run a correlation between Brady scores and crime rates for some time now, and I finally got around to doing it today. I did a bivariate correlation between a state's Brady score and its violent crime rate per 100,000 people and a state's Brady score and it's homicide rate per 100,000 people. I found that there is not a significant correlation between a state's Brady score and either its homicide rate or its violent crime rate.

This was a pretty crude comparison, and it's possible the results would have been different if I would have put more time into it. For example, obviously not every homicide or violent crime involves the use of a gun. Also, I don't believe that the state level is the best way to study gun crime, but I was more interested in doing it quickly. The crime rate in a large metropolitan area is going to be much different than the crime rate in a small rural town of a few hundred people, no matter what the gun laws are in a state.

Anyways, just thought I'd share this with everyone.

 

[denton]

Brady did something rather sneaky with their grading data: They left out DC and Puerto Rico, which are admittedly not states, but are state-like and in the FBI crime database. Both have draconian gun laws and extremely high rates of violent crime and homicide.

With those in the mix, I'd bet you'd get a negative correlation, that is tighter gun laws correlate with higher crime.

 

[avs11054]

I noticed that as well. I started listing the states by the FBI data (which included both of those) and then took them out when I noticed that Brady did not score either of those places. I'm not sure if it would have changed it that much though, the plots were all over the map. But both of those are defintely two places that would have added to the high brady score/high crime rate places.

 

[denton]

Regression is highly influenced by outliers and Puerto Rico is one. I think Puerto Rico alone might tip the balance from "influence indistinguishable from random noise" to "looser laws, lower crime".

PM me your email, and I'll send you a demo copy of a stats program that will change the way you do that stuff.

 

[Shadow 7D]

Also remember that percapata base, the majority of the population lives in draconian laws (NY, CA, MA etc.)

 

[chris in va]

I would think there also has to be a further breakdown of 'homicide' and 'justifiable homicide' as I suspect the two are lumped together.

 

[Solo]

Maybe in the Brady campaign's statistics, but in the FBI Uniform Crime Report and CDC reports on causes of death in America, they are separate categories.

 

[avs11054]

I would think there also has to be a further breakdown of 'homicide' and 'justifiable homicide' as I suspect the two are lumped together.

Maybe, but Im not totally sure. My stats for homicide came from the fbi's website. On there it is classified as murder and non-negligent manslaughter. Since they word it like that, it almost sounds as if they take out justified homicides.