Armed Correlations

[SecularMotion]

Well, to paraphrase a great Republican, here we go again. The details of Adam Lanza’s home environment—the armory of weapons there, the copy of the “NRA Guide to the Basics of Pistol Shooting,” there was a useful book kept close—are scarcely out before the insistence that there’s nothing, absolutely nothing, to see or to do returns, at a higher volume. The President talks, in his calm and conciliatory tones, about minimal gun control—that there’s no threat to responsible gun owners, just common sense—and gets, in return, no response, no counterproposal at all, just the usual toxic cocktail of fatalism and scorn. And he gets contemptuous references to his merely “emotional appeals” on the issue, to his talk, on Thursday, of “shame on us.” As though the horror of children ripped apart by a hundred and fifty-four bullets fired in less than five minutes is not itself rational evidence for change, as though unbearable parental grief is not itself an argument for altering the circumstance that made the mourning happen.

Then there is the “Well, we don’t really know what works!” faction, shrugging off talk of even those minimal measures. You know, there really isn’t much evidence that gun control reduces gun violence. The good and worthy David Brooks—whose role in life, unfortunately, seems to be to try to put a rational face on an irrational political faction—advises in a column that since gun deaths didn’t really end with the last, brief, assault-weapons ban, we shouldn’t press for another, at least not too hard.

Actually, it’s hard to find a more robust correlation in the social sciences than the one between gun laws and gun violence. The cry comes back: “But those are just correlations. They don’t prove causes!” And, indeed, the most recent damning study, published in that cranky, left-wing rag the Journal of the American Medical Association—which shows a clear correlation, state to state, between strong gun laws and less gun violence—ends with the orthodox injunction that the study could not alone determine cause-and-effect relationships, and that further studies are needed.

But when a scientific study ends by stating that there’s uncertainty about whether a correlation proves a cause, it doesn’t mean that correlations are meaningless in every circumstance. Everyone knows that creating false correlations between two unrelated elements is easy. But it can be that a correlation is so powerful and reliable that it may actually point to that rare thing in the social sciences, a demonstrable causal relation. As a wise man once said, “Correlation is not causation, but it sure is a hint.” When you can separate out a truly robust correlation between two elements in our social life, it’s a big deal.

http://www.newyorker.com/online/blogs/comment/2013/03/armed-correlations-gun-ownership-and-violence.html#ixzz2PEWbuKUx

[rrneck]

1. Of course correlation is important.

It gives the bloviaters something to bloviate about.

[DanTex]

4. This post is a great example of what I was talking about in my reply below.

Notice the complete lack of even an attempt to discuss the empirical evidence or form a cogent argument.

[rrneck]

5. So is the OP. nt

[ellisonz]

8. ^^^^

Blocked - And back to ignore you go.

[patrice]

2. K&R

[DanTex]

3. Good OP. The "correlation is not causation" line is used by denialists of all kinds,

ranging from gun nuts, to global warming deniers, creationists, etc.

Invariably, the person attempting to deny science has not a clue about the various ways that statisticians and social scientists have devised in order to distinguish non-causal or confounded correlations with causal links.

Yes, it is true that, as a matter of theory, it is impossible to draw causal conclusions by simply observing a system. But if we were to stop right there, that would mean that we could never draw any causal conclusions in any social sciences -- since you can't do clinical trials or perform controlled experiments -- and so we should just abandon the who enterprise of trying to reason empirically about large observable systems. This whole notion is very appealing to people who (a) are consistently on the wrong side of empirical evidence and (b) tend to trust their "gut" rather than data to begin with. In other words, people like right-wingers and gun nuts.

When it comes to gun violence, there is plenty of evidence beyond just raw correlations. For example, higher rates of gun ownership correlate with higher murder rates, this is well established. Of course, one can argue plausibly that there are non-causal explanations -- maybe people buy more guns when crime rates are higher, to protect themselves. Or maybe the rural lifestyle results in more violence for some non-gun related reason. However, on closer inspection, these other hypotheses fail. First, gun ownership does not correlate to other kinds of violent crime besides murder, something that pokes a hole virtually all of the alternate explanations (e.g. if the high homicide rate in the US were due to Americans being inherently violent, then we would also have high rates of robbery, rape, etc., but we don't). Also, the increase in murder rates associated with increased gun ownership is due to higher gun homicides -- rates of non-gun homicide are not affected.

And so on.

As a rule, with no exceptions I can think of, the gun nuts who throw out the "causation/correlation" line never have an intelligent alternate hypothesis for the empirical evidence. It's just an attempt to deny the obvious with a catchphrase they learned in junior high.

[ellisonz]

7. What I was taught in my sociology courses was that while "is causation" is frowned upon

That a statement such as "the quantitative analysis indicates that the statistical significance of the correlations leads us to conclude that the probability that the relationship is causal is strong." As a history major, I of course found this reluctance of sociologists to assign specific and conclusive explanations for events to be rather chickenshit. In short, I understand why they're reluctant make a more conclusive statement but as the OP article notes in regard to the recent JAMA study, it "ends with the orthodox injunction that the study could not alone determine cause-and-effect relationships, and that further studies are needed."

I have not read that study, but in general, I think this failure to them further explore under what conditions cause-and-effect relationships could be determined or what "further studies" are needed lends itself to misinterpretation. IMHO sociology doesn't do a good job of exploring and relating to its limits, whereas historians are going to more actively consider the probability that say perhaps a document might be uncovered that would disprove an argument based on the review of archival materials. Rare is the discovery of any new finding that would disprove or cause substantial reconsideration of a thesis. As someone with a degree in history, my thesis on this topic might be something like: "The history of gun ownership in modern time strongly illustrates the point that increasing small arms proliferation causes an increasing prevalence of small arms violence in society."

It's a simple and self-evident argument and doesn't lend itself to ready disabuse like sociologists are subject to, and I say this being readily aware of the bifurcated nature of sociologists between those who lean toward quantitative analysis and qualitative analysis. This is not to say that history isn't subject to abuse, I'm just saying that for lay purposes sociological explanation is more confusing to the average layman than historical explanation, and that by itself is an interesting question. I would speculate that what convinces is the mode of presentation of the "self-evident" argument that the OP author considers: "Common sense confirms what social science correlates. The United States, after all, is hardly the only rich country in the world with laws. American insularity and the ignorance of others is powerful, but it need not be quite so absolute."

I should probably bite my tongue now.

[jimmy the one]

9. CCCC hypothesis

dan: As a rule, with no exceptions I can think of, the gun nuts who throw out the "causation/correlation" line never have an intelligent alternate hypothesis for the empirical evidence. It's just an attempt to deny the obvious with a catchphrase they learned in junior high.

Agree mostly, since what they do throw out, like lott's ccw book 'more guns less crime', is generally so transparently pro gun propaganda, and thus not an intelligent alternate counter.

The axiom does not mean - 'causation can not follow from correlation', but that's how the rightwing generally likes to portray it when gun studies show a correlation between guns & more bad. As if the axiom itself proved 'more guns = less crime'.
More aptly - Causation can possibly follow from correlation'.

--------------------------------------------------------------------------------

ellisonz .. that while "is causation" is frowned upon .. a statement such as "the quantitative analysis indicates that the statistical significance of the correlations leads us to conclude that the probability that the relationship is causal is strong."

I usually see a correlation as being a corroboration towards causation. The more studies or examples you see with the same correlation, the more corroboration you have, & the greater the chance that there indeed is, a causal effect.

jto's CCCC hypothesis:'Reputable Corroborative Correlation increases the Chance of Causation'

But will it matter, ever? Gunworld would never accept any gun study which exposed the negative aspects of guns, they'd call it liberal junk science, &/or trash the source. Or if they couldn't do something like that they'd justify any negative aspect of guns with the old 'tree of liberty must be watered now & then', etc.... Rightwing concoctions are a dime a dozen.
David Hemenway of harvard & his JAMA gun(control) studies which are almost always well done & logical, get called biased & rigged by the pro gun crowd - the same crowd that disbelieves in global warming as a liberal scare tactic - like they'll ever have to live to see if it's true or not.
So we can turn out dozens of guncontrol studies & gain little headway, since they'd all get trashed or labelled junk, or 'peer reviewed' by rightwing statisticians as faulty or fraudulent, to get gobbled up by rightwing cretinism eager to have anything at all to counter liberals, the truth mattering not.

[DanTex]

11. You are right that the gun nuts will dismiss any scientific evidence, regardless of the quality.

In fact, this is a general tendency of conservatives in all areas, not just guns.

Regarding John Lott, that is an interesting story. True, at this point Lott is transparent pro-gun propaganda, but it wasn't always that way. At first, his research was taken seriously. But then the problems began to show up. His statistical methods were brittle and sensitive to changes in assumptions or specifications. His results weren't able to be replicated with extended data. And then of course he went off the rails, some data coding "errors" were found in his subsequent work, he invented an internet sock puppet and went around the internet praising himself, he fabricated a statistic in a major op-ed, etc.

[DanTex]

10. Part of the problem is communication to an unsophisticated audience.

Social scientists are technically correct to put in the caveats about causation. The thing is, among scientifically educated people, this is OK, because everyone understands the subtleties. For example, the fact that nothing can be absolutely "proven" by statistics, that causation can not be determined from observational studies, and so on. And yet, despite these caveats, it is still possible to observe the world and draw valid conclusions.

The problem arises with people with strong pre-existing beliefs and very little scientific education. People for whom everything must be black or white: either a study definitively proves a causal link, or it proves absolutely nothing. People who really aren't able or willing to look at the totality of the evidence, and use it to evaluate different hypotheses.

Of course, there is another side to this, where some people read the press release from a study that says "X is linked with Y" and come away thinking that it has been "proved" that X causes Y. That is also wrong. But what you see from denialists is a kind of pseudointellectual backlash, where they simply ignore the studies, and just go on believing what their "gut" tells them, satisfied that some slogan like "correlation is not causation" or "ivory tower elitism" allows them to dismiss the empirical evidence.

Also, the standards for scientific proof in hard sciences often aren't appropriate in social sciences, where it's hard to have repeatable controlled experiments. This is related to your point about the drawbacks of quantitative techniques in social sciences, versus the more qualitative approach of historians.

The thing is, we still need to make public policy decisions, so they have to be based on the best available evidence. We can't hope measure the effect of gun proliferation on homicide rates with the same precision that we measure the speed of light, but we can be reasonably confident that there is a substantial effect, certainly confident enough to warrant doing something about it.

[jimmy the one]

12. mary rosh

Dan Tex: True, at this point Lott is transparent pro-gun propaganda, but it wasn't always that way. At first, his research was taken seriously. But then the problems began to show up. His statistical methods were brittle and sensitive to changes in assumptions or specifications. His results weren't able to be replicated with extended data. And then of course he went off the rails, some data coding "errors" were found in his subsequent work, he invented an internet sock puppet and went around the internet praising himself, he fabricated a statistic in a major op-ed, etc.

I'm not sure when you speak of 'wasn't always that way', certainly not 'always' of course, but I first pegged him as a propagandist when he wrote 'more guns less crime' which was funded by olin foundation, a subsidiary of winchester ammo, & tried to play people for stupid by claiming his funding had nothing to do with winchester ammo or the gun lobby. Not directly - they're not stupid to put winchester name on a gun study - but the money funnelled in from winchester to olin funnelled into lott.
Lott also claimed an excuse similar to what happened to bellesiles (arming america, fine book) when an emory university hall flooded & damaged a lot of bell's notes; it was verified that the 'flood' happened, & his story believeable. But for lott not really believeable, he somehow was negligent in supplying data (I think what you allude to above as fabricated), and blamed it on failure of his hard drive I think, or something sounding much like a tapdance than a true story.

Yes, well familiar with mary rosh (lott's screename to promote himself & his books, largely on amazon I believe). "I'm the best professor I ever had!"
Sad, bellesiles came out of it disgraced, while lott seems to have suffered little, & even gained credibility from gunnuts, by his indiscretions.

[lastlib]

6. Kick & Rec!

pity too many RW pols can't read--that takes "edjimacashun"! Edjimacashun BAD!