Thread: States Loosen Concealed Carry Laws, Stir Debate

Originally Posted by Friedman

Nope all these assume that what we have observed is going to happen in the future and thus make assumptions about the uniformity of nature.

You'd (and I say that loosely as you're just coming out with the script that's fed the fake libertarians confronted by inconvenient evidence) have a point if I was referring to something like VAR modeling. Whilst I would use such methods for SME planning purposes, I wouldn't use them here. I'd merely test theory (repeatedly). That testing enables us to break through the dogma and ideological constrained

Originally Posted by Reiver

You'd (and I say that loosely as you're just coming out with the script that's fed the fake libertarians confronted by inconvenient evidence) have a point if I was referring to something like VAR modeling. Whilst I would use such methods for SME planning purposes, I wouldn't use them here. I'd merely test theory (repeatedly). That testing enables us to break through the dogma and ideological constrained

Nope, just because something happens lots of times under tests is no indication that it will continue to do in the future. Are you even aware of the problem of induction?

Originally Posted by Friedman

Nope, just because something happens lots of times under tests is no indication that it will continue to do in the future.

Take the dummy variable problem discussed earlier. Now we could clearly make the wrong policy choice because of a poor empirical specification. Hypothesis testing, however, confirms that right to carry effects are endogenous and dependent on county specific factors. Will that endogeneity simply disappear because of fake libertarian cliche? Nope. We could, if we wanted, run the regressions every year. Result? No change, except for wasted computer time

Originally Posted by Reiver

Take the dummy variable problem discussed earlier. Now we could clearly make the wrong policy choice because of a poor empirical specification. Hypothesis testing, however, confirms that right to carry effects are endogenous and dependent on county specific factors. Will that endogeneity simply disappear because of fake libertarian cliche? Nope. We could, if we wanted, run the regressions every year. Result? No change, except for wasted computer time

Irrelevant, was David Hume a 'fake libertarian'? You show no understanding of the problem of induction. Why do you think that the future will resemble the past?

Originally Posted by Friedman

You show no understanding of the problem of induction. Why do you think that the future will resemble the past?

No, you show no understanding of the empirical approach. The real problem, as illustrated in the previous post, is the use of ad hoc procedures to data mine a desired significant result. That is avoided through the use of theory to determine empirical specification. The analysis is then not a prediction of the future (a ludicrous idea as we have numerous variables, including interactive variables, changing. The clue is in the name). It is a direct test of the relevance of the theory

Originally Posted by Reiver

I'm well aware of the use of dummy variables in empirical research. Typically its not a problem. If we wanted to have a very simple gender control, bunging in a male dummy probably will suffice. Here, however, it ensures bias. If you look at the expression that they model you'll notice that its reliant on the right to carry effect being identical across all counties and independent of all county characteristics. Are you seriously going to suggest that those assumptions are realistic?

I am going to seriously assert that no arguments have been presented that would invalidate or present rebuttal to the study done by the University of Chicago. The claim that bias existed in the study is unsupported and no rebuttal studies have been presented to support the allegation. An allegation of inherent bias if unsupported by direct rebuttal by a compentent study remains nothing but an unsupported allegation.

The study doesn't even attempt to claim that all counties are the same but instead uses analysis of numerous counties to nullify the differences between them except as it related to concealed weapons. Random selection ensures an unbaised result albeit with a margin of error which is inherent in all statistical analysis. Statistical analysis never presents an absolute as even a 1% margin of error implies that the odds against the study results being correct are 99:1. It could be logically argued that the margin of error, inspite of the odds against it, could misrepresent reality but it takes the position of "long odds" that would have to be supported by valid argument with a statistical analysis that reflect that random chance used by a study resulted in invalid results.

So where is the statistical analysis that directly refutes the findings by the University of Chicago? I'm waiting for that and will reject any simple unsupported allegations that condemn a form of statistical analysis that has been shown to be very accurate in general usage. Prove that the analysis is in error.

Originally Posted by Shiva_TD

I am going to seriously assert that no arguments have been presented that would invalidate or present rebuttal to the study done by the University of Chicago.

You surely know that, to assume that the right to carry effect is identical across all counties and independent of all county characteristics, is on a par with assuming capitalism and perfect competition co-incident. Its a ludicrous assumption.

The claim that bias existed in the study is unsupported and no rebuttal studies have been presented to support the allegation.

Actually I referred to how the paper has been subsequently dismissed. I'm happy to give a reference in support: Rubin and Dezhbakhsh (2003, The effect of concealed handgun laws on crime: beyond the dummy variables, International Review of Law & Economics, Vol. 23, pp 199-217)

The study doesn't even attempt to claim that all counties are the same but instead uses analysis of numerous counties to nullify the differences between them except as it related to concealed weapons.

The study has to assume all counties behave the same or the dummy variable will be biased. There's no debate in that.

Originally Posted by Reiver

You surely know that, to assume that the right to carry effect is identical across all counties and independent of all county characteristics, is on a par with assuming capitalism and perfect competition co-incident. Its a ludicrous assumption.


Actually I referred to how the paper has been subsequently dismissed. I'm happy to give a reference in support: Rubin and Dezhbakhsh (2003, The effect of concealed handgun laws on crime: beyond the dummy variables, International Review of Law & Economics, Vol. 23, pp 199-217)


The study has to assume all counties behave the same or the dummy variable will be biased. There's no debate in that.

Your not aware that Rubin made a mistake. He didn't renormalize the psi function against background variables from other countries. The end result is a highly skewed against concealed carry.

Originally Posted by Reiver

You surely know that, to assume that the right to carry effect is identical across all counties and independent of all county characteristics, is on a par with assuming capitalism and perfect competition co-incident. Its a ludicrous assumption.


Actually I referred to how the paper has been subsequently dismissed. I'm happy to give a reference in support: Rubin and Dezhbakhsh (2003, The effect of concealed handgun laws on crime: beyond the dummy variables, International Review of Law & Economics, Vol. 23, pp 199-217)


The study has to assume all counties behave the same or the dummy variable will be biased. There's no debate in that.

The study does not claim that all counties would be affected identically but instead establishes that as a total the average affect would be realized.

Is there a link to the rebuttal argument or is this some abscure paper that cannot be referenced online. Of note the "International Review of Law & Economics" is a British review that addresses international conditions while the University of Chicago study was expressly limited to the United States.

Talk about introducting bias nothing would create more bias that trying to apply the results of an international study in addressing a purely American situation. The differences in the nations of the world are far greater than the differences in counties in the United States. The very structure of government is different between the United States and any other government as is the foundation for government where the US is the only nation where the protections of the inalienable (unalienable) Rights of the People was established as the primary role of our government. No other nation has this foundation for their government.

Of note capitalism is not based upon "perfect" competition and that has always been a failed argument of the socialists. It merely establishes that a free market where the Rights of the Individual are protected, including the Right of Property, and the enforcement of contract law are superior to government interventionism in the ecomony. The current US recession is a perfect example of government interventionism in the economy resulting in economic disaster and the failure of the US government to protect the Rights of the People as well as it's blatant failure to enforce contract law.

Originally Posted by Shiva_TD

The study does not claim that all counties would be affected identically but instead establishes that as a total the average affect would be realized.

You misunderstand. The dummy variable approach, by definition, claims that.

Is there a link to the rebuttal argument or is this some abscure paper that cannot be referenced online.

The paper is here.

Of note the "International Review of Law & Economics" is a British review that addresses international conditions while the University of Chicago study was expressly limited to the United States.

The paper uses Lott and Mustard's data, so you couldn't be more wrong.

Talk about introducting bias nothing would create more bias that trying to apply the results of an international study in addressing a purely American situation.

We have something simple here. We have a well known problem with dummy variables. We have a peer reviewed paper that finds that, once this methodological flaw is eliminated, the results are significantly affected. You may not like those results, as they attack your original claim, but there's not much we can do about that.