Mass shooting in the middle school Vladislav Ribnikar in Belgrade

[FlatAssembler]

It has been shown time and time again, in various places around the world, that having restrictions on who should own firearms (and enforcing those restrictions) reduces gun violence.

I think that even if you discard the Gary Kleck's study (and similar studies), that is still highly unlikely. Like ReasonTV put it, there are around 400 million guns in the US, and the vast majority of gun control laws only affect the sales of new guns. So gun control laws can at best reduce the total amount of guns by around 1%. And the incidence of homicide varies by around 6% from one year to the next. No real study can detect the effect of the gun control laws. The signal-to-noise ratio is just too low.

In fact, there is a direct correlation between how restrictive gun laws are, how well they're enforced, and incidents of gun violence.

I am not sure about that. The UK has much more strict gun control laws than Serbia and Croatia do, yet the UK is much higher in violent crime than Croatia and Serbia are.
And even if that's true, I don't think that suggests gun control works. Think of it this way: if you lived in a high-crime city, wouldn't you want to get a gun to defend yourself? Maybe it's the people in high-crime areas using guns to defend themselves, rather than guns causing areas to become high in crime.

This is so glaringly true that it can't be sensibly debated.

To me it seems like you can just as easily argue for the opposite. An undeniable truth is that most mass shootings happen in gun-free zones. Schools are gun-free zones. Social security offices are even more strictly enforced gun-free zones than schools are. Yet the mass shooting in Đakovo in 2019 happened right in a social security office. There was an armed police officer at the door, and the mass shooter Andrija Drežnjak passed by him. Could it be that mass shooters perceive gun-free zones as "soft targets"?
Far-fetched, yes, but not more far-fetched than the idea that gun control somehow prevents mass shootings.

We've already discussed the problems with Kleck's study and methodology. You need to stop bringing that up.

Then what are those critical problems with the Kleck's study, that you can reject it without pointing to a better study?

If you are going to say "too small sample size", then you will simply prove you have no idea how science works.
Sometimes a study with a sample size of one is enough to prove a scientific hypothesis wrong. In fact, sometimes even one measurement is. I once thought that, since the English language has 21 consonants, the collision entropy of the consonants in a long text on the English language will be around log2(21) bits per symbol. Only one measurement was enough to prove that hypothesis wrong, a measurement showing that the collision entropy of the consonants in a long text on the English language is around log2(11) bits per symbol. Then I formulated a scientific hypothesis that most of the collision entropy of a language goes to the syntax, and that, if we eliminate the syntax, the collision entropy of the consonants will be around log2(21) bits per symbol. Once again, a study with a sample size of one (only the English language) is enough to prove that hypothesis wrong, as the collision entropy of the consonants in the Aspell dictionary of the English language (so, a text without a syntax) is only log2(13) bits per symbol.
Then I formulated a scientific hypothesis that the collision entropy of the syntax (the difference between the collision entropy of a long text and the Aspell word-list) is strictly correlated with the complexity of the syntax of the language. So I did a few more measurements. Once I had a sample of only three languages (Croatian, English and German), I was able to prove my hypothesis wrong. Namely, I knew (because I studied both English and German for years in school) that English has a more complicated syntax than German, yet my measurements showed that German has higher collision entropy of the syntax than English does.
Then I made measurements for two more languages, French and Italian, and I made a table with the data for 5 languages: English ("Engleski"), Croatian ("Hrvatski"), German ("Njemački"), Italian ("Talijanski") and French ("Francuski"):

Then I noticed that English and French have the lowest collision entropy in both a long text and the Aspell word list. What else do those two languages have in common? Well, they have a much deeper orthography than the other languages in the table. What is the probability of that happening if we assume the depth of the orthography has no effect on the collision entropy of the language? Well, the probability of that happening (the p-value of my observation) is 1/((5!/(2!*(5-2)!))^2)*2=1/50. P-values below 1/20 are usually considered statistically significant. So, a study with the sample size of 5 has a statistically significant result. It suggests that a deep orthography decreases the collision entropy of a language. Does it prove that beyond reasonable doubt? Well, I don't think so, since that is a pretty extraordinary claim, but I think it strongly suggests that.
I think that my latest paper (published in Valpovački Godišnjak and Regionalne Studije, and from which that table is taken) and the Gary Kleck's study are examples of good social science.
Gary Kleck's study has a bigger sample size than any study comparing languages can possibly have, because there are only around 5'000 languages around the world.

And if you are going to say "many people are delusional that a gun saved their life", keep in mind that Gary Kleck asked the participants "Have you actually seen the attacker?", and didn't take those who haven't actually seen the attacker seriously. Gary Kleck's study might even be an underestimate, rather than a 10x overestimate (which is necessarily for a gun prohibition to reduce violent crime).

Experience has shown that when mentally ill people get their hands on a firearm, they don't use it for self-defense. They use it to shoot up a school or an office or the family that asked them to stop shooting so their baby could sleep.

Having a diagnosed mental illness is not a good predictor of being a mass shooter. Some studies even show that people with a diagnosed mental illness are less likely to be a mass shooter than other people are. My guess is that it is because mass shooters aren't seeking help, they are convinced they are mentally healthy.