Kevin P sums it up

Re: this bogus "new" Harvard study that tries to claim that more guns automatically equal more crime, Kevin P. leaves a comment in Deltoid that reflects my thoughts on the whole issue:

I have a Master's degree in Chemical Engineering and work in the semiconductor industry. As a process engineer, I routinely used statistically designed experiments to improve process and chip performance. I rapidly learned that you can easily play with your statistical models to tell you whatever you would like to hear.

In my line of work, it was easy to verify the model by running a new experiment on actual living and breathing wafers. This kept everyone honest.

This unfortunately does not apply to the social sciences. You (usually) cannot run a verification experiment to prove that your model was correct. In this particular case, that would involve actually enumerating gun ownership by visiting and searching every private home in the target area. This is of course impossible - you can't search even a single home without a warrant. The study then is unverifiable. I suspect that most social science is like this but we don't hear much about it because many of these studies may be on uncontroversial subjects.

But this is of course on the subject of gun control. There is a systematic effort, funded by the Joyce Foundation in recent years to the tune of millions of dollars to investigate the private lawful ownership of guns contributes to crime. Coincidentally, almost all of these studies find that private lawful gun ownership contributes to crime, and recommends various gun control methods. The entire field has become poisoned. While it is possible to find honest scholarship in this area, Joyce Foundation studies are not the place to go looking for them.

Furthermore, if a study contradicts real world experience, it must be able to explain the contradiction in a compelling way. In my real world experience, the gun culture is composed of tens of millions of law abiding Americans who take responsible care of their firearms. A very small number of criminals, concentrated in geographic and demographic areas are responsible for the vast majority of crime. Studies conducted in the liberal bubble of Harvard, far removed from the gun culture of America lack any kind of real world experience or perspective. Unsurprisingly, their findings mirror their authors' prejudices (and their funders' desires).

Real science will get your Boyle's law, which is applicable throughout the universe. So-called "gun violence research" is not "science" because you can always find examples that contradict the supposed rule. EX: anti-gun-owner New Jersey has the statistically identical homicide rate of pro-gun-owner Florida, even though Florida has a higher poverty-ridden population, and has more people living in urbanized areas than are in the entire state of New Jersey.