Sorry, I hadn’t read your initial post in this thread. I was addressing only what you wrote in your one comment.
To address the Mother Jones data, I would say that you are probably reading it wrong.
- A desire to murder and the ability to murder are completely unrelated to each other. Your average murderer is caught without ever having done anything, because he’s a dumbass. Occasionally, though, he kills a dozen or more, because he has the intellect, demeanor, and free time to pull it off. A metric that is based on the sheer number of people killed is going to swing massively just based on the capability of the psychos that year. It’s pretty much just random.
- Or, if the metric we look at is the total number of cases in a year (rather than the total number of deaths), then you have note that we’re talking about a handful per year in a country with 300,000,000 people. If there’s a 2 in 100,000,000 chance that someone would try to execute a mass murder in a year, on average we would expect about 6 per year. That doesn’t mean that you actually get 6 every year though. You would probably expect it to fluctuate between 0 and 20, randomly. Randomness is a lot more friendly to patterns than our brain likes so we tend to find patterns in small datasets that aren’t there. It’s entirely reasonable to get 5 or 6 years in a row with low numbers and then another five with high numbers, and there’s genuinely no explanation for that divide other than random happenstance. A good video: https://youtu.be/tP-Ipsat90c
- I suspect that a large artifact in their data is the availability of information to them. With computers, police stations can record and find cases that match a certain search query. Probably most police stations weren’t computerized until the mid to late 90s. With the Internet, it becomes easier for journalists to find more cases. Between those two things, we would expect to see more homicides that conform to their criteria as we move ahead in time, just on the basis of the ease of discovery. I strongly suspect that their data is missing a majority of cases that have occurred during the time period they searched.
- The choice of a gun as the implement of murder is arbitrary. Guns are largely unrelated to the homicide rate and usually reduce it, so singling it out as a metric is silly. Why not just document “mass murder”? Why, instead, document “mass murder with a gun”? To be sure, if you keep track of mass murderers and the implements that they use, it doesn’t hurt to track the implement they use and - when there’s enough data to make a conclusion - see if it’s likely that you could reduce the rate by changing the rules around access and use of that implement (if we ignore the 2nd Amendment). But that’s at the end of the scientific processes. Our largest and most effective mass murderers used an airplane. We have had effective mass murderers use swords, arson, bombs, and other means to accomplish the same thing. I wouldn’t be surprised if arson is the most effective and common means of mass homicide being utilized today so, while the FBI probably has people scanning Facebook posts to find people talking about guns, they potentially aren’t having it looking for accelerants simply because the political focus is elsewhere, and that could be allowing people to die needlessly.
- Legally, not a lot has changed since 1982. We can’t expect any regulatory effects on the mass homicide by gun rate. The Internet and social media is, plausibly, a cause of increased mass homicide rates starting from the late 90s. But, at the same time, homicide rates are down since 1982 and that’s popularly hypothesized to be due to a reduction in lead fumes making people crazy and violent. It seems strange that general craziness would reduce and yet mass homicides would purely increase. Even if we accept that the Internet taps in to some murderous desire and unlocks it even better than lead does, we should see a reduction in mass homicide until the greater take-up of the internet. The dataset should look like a V not a ramp. This strongly implies that the data is incomplete.
- I haven’t checked the growth of the US population, but I doubt that it was 0 in 1975, so I also doubt the accuracy of the general trend of the data on that basis.
I would say that the data is probably a reasonable set of examples of mass murder. But in terms of spotting trends, the numbers are too small to be statistically significant, the data collection seems suspect, and I question the validity or usefulness of the “gun” criteria when it comes to the ultimate question of trying to save the most lives that can be saved.