[SauronOfMordor]

Hence if you examined the relationship between CC and crime without controling for crack addiction, you would that CC tends reduce crime even though it may not. This is called correlated omitted variable bias.

The relevant stat is not the level of crime in CCW states, but the level of crime COMMITTED BY CCW HOLDERS (which is generally insignificant). The only justification for restricting CCW is if CCW holders were likely to commit crimes with their guns. Since the gun-grabbers are unable to show anything like this, they have to work with other statistics

[traditionalist]

The relevant stat is not the level of crime in CCW states, but the level of crime COMMITTED BY CCW HOLDERS (which is generally insignificant). The only justification for restricting CCW is if CCW holders were likely to commit crimes with their guns.

Oh, I agree with you, but John Lott's claims were that CCW was associated with lower crime trends and levels. All Donohue's paper shows, using methodology I find sound, is that Lott's original result is not robust.

In his research paper, Donohue does not claim that his results justify restricting CCW. He only does that in the LA times article, but his statements are clearly not supported by his own findings.