We aren't trying to evaluate correlation or causation. We are simply determining if the hypothesis that "California's strict gun control laws decreased California's homicides" is or is not a true statement. just as if we were trying to determine if our hypothetical blood pressure pill lowers blood pressure. It either does or it doesn't. Furthermore, we don't give one single fart if it decreases blood pressure in populations that don't have high blood pressure to begin with which is what you're looking at when you compare statewide per 100,000 homicide rates. Homicide and gun violence are not problems in rural areas. Including the entire statewide population in your analysis is like including white people in your sickle cell anemia prevention medical trial. There's no point. It only leads to a bad conclusion. It doesn't take a rocket scientist to figure this out. I don't care what your degree is in, you're wrong here.
"California's strict gun control laws decreased California's homicides"
Please read that again for yourself, carefully. Notice it is a causal statement, "decreased" being the operative term. In hypothesis testing, you still need to rule out confounding variables.
just as if we were trying to determine if our hypothetical blood pressure pill lowers blood pressure. It either does or it doesn't.
Nope, not that simple. It does or it doesn't, but to actually know that you have to control for confounding factors. If you give the pill to people that then also go out and smoke 3 packs a day and their blood pressure is not lower you cannot automatically attribute that to the effect of the drug not being there, because its possible that it was and the effect was masked by the smoking.
Furthermore, we don't give one single fart if it decreases blood pressure in populations that don't have high blood pressure to begin with which is what you're looking at when you compare statewide per 100,000 homicide rates.
This makes absolutely no sense. If there are no homicides then the rate must be identically zero. Giving a rate per population does not remove the fact that homicides occurred, it simply accounts for the fact that with more people you would expect more total homicides.
Homicide and gun violence are not problems in rural areas.
That is a bit of a generalization, there is only a weak link, on a state level, between urban density and homicide.
Including the entire statewide population in your analysis is like including white people in your sickle cell anemia prevention medical trial. There's no point. It only leads to a bad conclusion.
That analogy is both incorrect and nonsensical. The total count of homicides still includes the entire state as well, so that does not change anything even if you did use it. But furthermore rural areas have relatively few people in them, so they hardly change the rate anyway.
Finally, why does this even matter if you are comparing two states together? The impact of including the entire state applies to both.
Additionally, you continue to use medical trials as an example without realizing that not everything that is done in a medial trial applies to this sort of analysis. Medical trials have the ability to control for many variables in the design of the trial, but this is after the fact statistical analysis which does not have that luxury. Economists and others in related fields encounter this all the time, you have a natural experiment that occurs in the wild and need to try and tease the correct conclusions out of the data. That is a very different proposition than designing a double blind study where you can carefully control most variables, and get accurate data on the rest.
To your point, obviously if you were going to design a sickle cell trial you would want people that have the disease. But notice, every state has homicides so there is no rational basis for excluding any of them. Nor is there a rational basis for using total homicides, which completely fails to control for state population, and trying to use that to draw any meaningful conclusions.
It doesn't take a rocket scientist to figure this out. I don't care what your degree is in, you're wrong here.
Note that I never did mention my degree, because I don't really like appeals to authority, they are something of a logical fallacy. That said, at least one engineer has chimed in to point out that this analysis is crap, and for what it is worth my background is in mathematics, economics, and statistics.
Your argument here is almost comically unsound, and would be handily refuted by anyone with even a minimal understanding of how statistics or math works. It would fail to convince anyone that is not already pro 2A, and frankly it does not convince some of us who are pro 2A anyway.
