Washington Post:

Stanford law professor John Donohue and his colleagues have added another full decade to the analysis, extending it through 2010, and have concluded that the opposite of Lott and Mustard’s original conclusion is true: more guns equal more crime.

“The totality of the evidence based on educated judgments about the best statistical models suggests that right-to-carry laws are associated with substantially higher rates” of aggravated assault, robbery, rape and murder, Donohue said in an interview with the Stanford Report. The evidence suggests that right-to-carry laws are associated with an 8 percent increase in the incidence of aggravated assault, according to Donohue. He says this number is likely a floor, and that some statistical methods show an increase of 33 percent in aggravated assaults involving a firearm after the passage of right-to-carry laws.

These findings build on and strengthen the conclusions of Donohue’s earlier research, which only used data through 2006. In addition to having nearly two decades’ worth of additional data to work with, Donohue’s findings also improve upon Lott and Mustard’s research by using a variety of different statistical models, as well as controlling for a number of confounding factors, like the crack epidemic of the early 1990s.

These new findings are strong. But there’s rarely such a thing as a slam-dunk in social science research. Donohue notes that “different statistical models can yield different estimated effects, and our ability to ascertain the best model is imperfect.” Teasing out cause from effect in social science research is often a fraught proposition.

But for this very reason it’s important for policymakers on both sides of the gun control debate to exercise caution in interpreting the findings of any one study. Gun rights advocates have undoubtedly placed too much stock in Lott and Mustard’s original study, which is now going on 20 years old. The best policy is often informed by good research. And as researchers revisit their data and assumptions, it makes sense for policymakers to do the same.

Occasionally something comes along on which I’m uniquely qualified to comment.  I’ve explained before that I don’t like John Lott’s approach (here and in Holding Human Rights Hostage To Favorable Statistical Outcomes).  See also Kurt Hofmann on this same subject.

But it’s important to be able to discern science from pseudo- or non-science or bad science.  I work in science and engineering every day.  I have for 33 years of my career.  I am a registered professional engineer.  An example of bad science might be AGW (anthropogenic global warming).  The notion that a “researcher” can prove anything about trends by claiming 1 degree C change over a half a millennia is ludicrous on its face.  Furthermore, trusting tree ring data is only valuable if your thesis doesn’t suffer from falsification of data (i.e., the “hockey stick” lie).  But even if tree rings could be a trusted source of information when we have no recorded data, the information is statistically insignificant.  No one with whom I work, engineer or scientist, not one of the hundreds I know, would actually put his or her name on such a calculation or thesis, especially if it involved affixing a PE seal to the work.  AGW is bad science.

Now to what is actual science.  If I use a computer model of a system (which involves physical and engineering calculations) and generate a curve of results from input that has been perturbed, or in other words, a sensitivity study, and I generate a curve fit with TableCurve-2D, and then put that polynomial into MathCad and integrate to a solution (because for some reason I wanted the results from integration), that is science and engineering.

Or say that I use the Bernoulli equation and information on pipes from the Crane Flow of Fluids Technical Paper No. 410, or Cameron Hydraulic Data, to build a piping network, that is science and engineering.  Or say I want to evaluate the performance of a projectile and I use Newtonian physics and ignore aerodynamic drag for simplicity, or say that I do not ignore drag and I account for it, that is science and engineering.  Or finally, let’s say that I use Henry’s law to ascertain how much of a gas is dissolved in the liquid in a system, that is science and engineering.

The grand mistake in the article above is that it uses the phrase “social science.”  There is no such thing.  That’s a myth perpetrated by the sociologists and psychologists.  When you are dealing with humans who have choice and volition, there is no mathematical or physical model you can invoke in order to make it science.

I know what sociologists and psychologists are thinking right about now.  You are all behaviorists; man’s actions and choices are the outcome of syntactical impulses, chemical reactions, his history, or something of the sort.  And you so want it to be that you are scientists, and you so badly want for what you do be to considered science.  But you are not, and it is not.

The right reaction to articles such as this is to assert, and rightly so, that if I have a weapon and handle it with care and concern, train with it, am diligent to observe all the rules of safety and self defense, it is more likely that I will be able to defend myself and my family.  I am not a statistic.  I am not subject to the application of mean and standard deviation.  I am not part of the collective, and so it doesn’t matter what the collectivists want me to think about myself.

And don’t ever listen to someone who begins by telling you he is a “social scientist.”