I am struck by the lack of data in the gun-control conversation I have heard on TV and radio lately. Perhaps this is because, as my friend Ethan wrote, “the dialogue on gun laws in the US is locked between two positions that are both completely divorced from reality,” and there is no room for data in that kind of dialogue. Whatever the reason, casual listeners like myself hear a lot of talk and very little statistics.

What kind of guns fire the bullets that kill people in this country? If for some reason we’re only interested in mass killings (which most of the time seems to be the case), what kinds of guns fired those bullets? What percentage of the guns were purchased (new? used?) by the shooter? What percentage stolen? How many killings were “good guys” killing “bad guys”? In other words, how much of the killing that has happened would be have been effected by each set of proposed gun-control regulations?

What are the media diets of murderers? Do they differ in any real way from non-murderers? Is there any hint of a dose-response relationship between media violence/1st person shooters and murder? In other words, do you increase your chances of killing someone by playing Half-Life 2 for 900 hours?

I am in a similar state of frustrated ignorance about mental illness and mass killings, or murder in general. This is a subject that interests me greatly, as I work in the mental health system. It is somehow much more controversial to increase the regulations on guns than it is to create a national database of people who have been diagnosed with mental disorders. I can’t think of a better way to reduce the number of people who get help with psychological troubles that to create a database of them.

Maybe I could be swayed, though, if I had some facts to work with. How many murderers have been diagnosed with what disorders? What psych meds were they on? How many had inpatient vs outpatient treatment? How do those numbers compare to the general populations? Clinical populations?

One thing no one seems to talk about is that mass killings could be a trend the way methods of suicide have trends. I find this idea plausible and quite disturbing. Now, in the US, when you realize that you need to do something spectacularly evil, the obvious thing to do is go to a school or mall and kill a bunch of people. I went to school in the 1980s, before this trend established itself, and I feel lucky to have gotten out when I did. I remember at least one kid who was bullied so bad I’m surprised he didn’t bring a gun to school. It just wasn’t what you did yet.

From that perspective, it seems unlikely that making it more difficult to buy certain guns will make much of a difference. (Of course, I have no empirical evidence to back myself up there, and I am happy to be swayed by evidence.) From that perspective, the most and perhaps only effective intervention for the problem would be the media refusing to report the incidents. This seems way less likely than gun-control legislation, and media-control legislation is more obviously unconstitutional.

The question of constitutionality of gun-control is also confusing to me. Assuming “arms” was synonymous with “weapons” in 1791, there was a pretty good constitutional argument against any kind of weapons-control 200 years ago. That argument is long obsolete. I don’t hear anyone advocating unregulated access to rocket launchers or nuclear weapons. So we’re left in this zone that is not mentioned by the constitution, where we have to draw a line between weapons we want to regulate and those that we don’t using the democratic process. That process doesn’t seem to be about the constitution any more, except when gun-rights advocates invoke it without getting into the issue of rocket launchers or Adam Lanza with a combat drone.

Finally, I am most baffled by the folks who are talking about “starting another civil war” if the government tries to “take our guns.” It’s not just that no one of any consequence is talking about disarming anyone. It’s the most clear example of the pro-military-anti-government disconnect in the right wing. Just like conservatives never seem to be thinking of the military when they mention government spending or government employees, they must not be thinking of our military as “the government” when they talk about a civil war. A civil war would not be fought against corrupt, middle-aged bureaucrats and politicians, it would be fought and lost against the United States military.


This is a cool visual presentation of global health and wealth. I often find presentations of data to be either dense and non-intuitive or boring. This one is interesting and inspiring.

Hans Rosling also has a TED talk here that is really worth watching, on population, fertility, child survival rates, and wealth.

(By the way, I found this clip while looking for good map animations, which seem like a great way to present data. It’s pretty slow going so far, though. Any recommendations?)

In some ways it is nice that psychology research is fed to us in the discrete package that is the journal article: Each package can be edited into some version of readability and peer-reviewed for credibility.

That the process stops there, however, is an anachronism. It used to be that publishing the actual data interpreted in the article would have taken up too much space in paper journals, but on the internet it would be easy to do, and far more useful than just the analysis.

Imagine being able to go back in and re-run the statistics for an experiment, or try out other analyses–especially while analyses that throw away information like that old standby, the median-split ANOVA, are still accepted by journals. Imagine how much more powerful meta-analyses could be if it was standard practice to publish the data. Every research project would be a potential collaboration.

In fact, like scanning the Library of Congress, we could retrospectively publish all the data from every published article in the archives! What a resource that would be.

It might not work so well for qualitative research, the data of which are interviews with individuals whose confidentiality has to be protected. For quantitative research, though, it would be easy to protect anonymity. I see no downside except, I suppose, for researchers who are fudging their numbers.