May 31, 2013

"[I]n a hypothetical state with no guns and no religion, we’d expect a suicide rate of 17 (per 100,000)..."

"In a state where everyone had guns, and no one practiced religion, we’d expect a suicide rate of 39. If everyone were religious, but no guns: 11. Everyone religious, everyone owns a gun: 21. From a political perspective, there’s really something for both the left and right to like here."

58 comments:

The Drill SGT said...

The problem with assuming that guns double the suicide rate is that some forms of gunless suicide are easily conflated with accidents.

Expat(ish) said...

I read this out loud to my wife, who is a Ph.D. Sociologist, and she snorted.

Apparently there is a ton of scholarly research with actual, you know, scholars running numbers across cultures and using time series data.

I note that none of these so-called journalists seem to use any of that research.

-XC

Lem said...
This comment has been removed by the author.
Lem said...
This comment has been removed by the author.
Moose said...

The suicide rate in the the UK is 12 per 1000. The suicide rate in the US is 14 per 1000.

That pretty much negates the "guns cause suicides" argument.

SteveR said...

"we'd expect"?

Sounds very scientific

Marshal said...

I think it's interesting that the left attacks the right as authoritarian for arguing against assisted suicide, then uses suicide prevention as a reason to limit other constitutional rights. To square this philosophically you have to believe people have the right to suicide as long as it is not accomplished using tools you have a constitutional right to posess.

Or you can accept this isn't a principle but rather an argument they advance because it allows them to charge their opponents with supporting death.

Phil 3:14 said...

Its statistical correlation based on observational data; it can't demonstrate causation.

Having said that why does it surprise anyone that a higher prevalence of guns is correlated with a higher incidence of suicide? I'm assuming the definition used here is completion of suicide. Guns are a more effective tool for suicide.

I'm not a gun guy but this article doesn't lead me to conclude that stricter guns are needed to reduce suicide

Henry said...

Do we know the religiosity of the people who commit suicide or just the religiosity of the people prevalent in their environs?

Hypothetical states are very hard to move to.

phx said...

"we'd expect"?

Sounds very scientific


That's what a hypothesis sounds like.

Phil 3:14 said...

But I do conclude that all new gun owners should be required to attend church services for the first 4 weeks of ownership.

X said...

what would the murder rate be in a gunless society?

would women be less equal in a gunless society?

DADvocate said...

Hypothetical states never exist any more than utopian states exist. What we expect and what we get are often quite different.

Darrell said...

So having religious beliefs seems to be the most important factor then. Guns are just noise--used more where available, less where not.

madAsHell said...

No religion?

No religion???

There will always be a religion.
How do you explain global warming??

Henry said...

At least this regression is using data from the actual United States. That is far less annoying than the typical pseudo-science of comparing the U.S. to some small Nordic country and wondering why a motley jumble of WASPS, Irish, Germans, Italians, Africans, Chinese, East Indians, Spaniards, and their neighbors, don't behave like Anderssons.

David said...

We have had a gunless society.

The biggest meanest most agressive males prevailed, and then organized packs of males.



Jay Vogt said...

Well, whatever your politics are, this study runs into the problem of correlated independent variables at about 105 mph. Effect attribution is an exceedingly tricky exercise that almost always breaks down under any kind of thoughtful scrutiny.

ErnieG said...

"Oh, people can come up with statistics to prove anything. 14% of people know that."

--Homer Simpson

Jay Vogt said...

Or, as I've been cautioned a time or two, "there's nothing more ugly than seeing a beautiful theory ravaged by an ugly gang of facts".

SteveR said...

That's what a hypothesis sounds like.

Well maybe in that field of study. I was referring to science.

CWJ said...
This comment has been removed by the author.
CWJ said...

Multiple regression with a rate as the dependent variable is a tricky business. I'd have to to know exactly how the models were specified and whether any data transformations were used before I gave this any credence.

John said...

Anyone interested can look in the Statistical Abstract of the US, Section 30, Table 1345

It has suicide rates for about 30 countries. US is 10.1, OECD average is 11.1

Japan, no guns at all in civilian hands, has a rate of 19, Korea, 21. France, very few guns, 13.5 and so on.

Even our northern neighbor outdoes us by a tenth of a point and has a lower rate of gun ownership.

Of course stuff like that takes all the fun out of a good story.

I wonder how many journalists have even heard of the Statistical Abstract?

For those here who have not, it is an excellent first go-to for statistics on pretty much anything you might want. 2012 edition is here:

http://www.census.gov/compendia/statab/2012edition.html

John Henry

Jim said...

So I guess they really DID imagine there was no heaven; and more people offed themselves because of it.

The Godfather said...

There may be "something" in this study to please both liberals and conservatives, assuming they are easily fooled by stupid studies.

What "religion" are we talking about? Because the data is from the US, I assume it's predominantly Roman Catholic and Baptist (I recall seeing a map that color-coded each US state by which denomination was the largest, and you needed only those two colors). The Catholic Church strongly condemns suicide; Baptists I think condemn it but less strongly. If the "religion" were one that honored suicide, so long as you took some infidels with you, you'd probably get a very different result.

As for "no religion", I don't know where you find a community in the US with no religion, so don't see how you can study that.

And what about "with guns" or "without guns"? Chicago and New York City have very strict gun laws (although I assume that'll change as the Constitution eventually reaches those places). Are we to assume that they have very low suicide rates compared to, say Wyoming?

This is just BS.

Paul said...

Are they gonna ban bridges to (many jump off them each year?) Ban sleeping pills (many use them to commit suicide each year?) Ban ropes? Ban cars (how many run them into trees on purpose?) Ban razor blades? Knives?

The author of the study is just full of it and has no idea what they are taking about.

Reminds me of Belsides who lied about his sources for gun use by early colonialist. He faked the whole thing, wrote a book, and was caught lying.

phx said...
This comment has been removed by the author.
edutcher said...

From the linked article in the linked article:

"We must also remember that all we’re really finding here are correlations. Making the leap to causation is just that: a leap."

Religion provides spiritual comfort.

Firearms provide psychological and physical comfort.

Next question.

t-man said...

This post brought to you by John Lennon.

Capt. Schmoe said...

People who really want to commit suicide will. Guns or not, there are plenty of methods available to successfully complete the task.

The statistic equating the number of suicides to the availability of guns more likely implies that the firearm was easier or more convenient than other methods - not that if there was no gun the event would not have happened.

Suicide by train is somewhat popular around here, especially among the transient population.

Someone who can hunker down on the tracks, giving the world the finger while drowning out the sounds of the approaching locomotive is someone who really wants to die.



Christopher said...

Having said that why does it surprise anyone that a higher prevalence of guns is correlated with a higher incidence of suicide?

Because it isn't. See @John above.

TomHynes said...

Why is gun suicide a bad thing? Don't most progressives and libertarians support a right to die? Should you only have a right do die in a government approved fashion - assisted suicide - and self help is wrong?

Rusty said...

The "hypothetical state" rarely survives its contact with reality.

Rusty said...

TomHynes said...
Why is gun suicide a bad thing? Don't most progressives and libertarians support a right to die? Should you only have a right do die in a government approved fashion - assisted suicide - and self help is wrong?

Apparently it's fine when your no longer of any use to the state.

Certainly puts the ACA in perspective, doesn't it.

EMD said...

This would be a problem if I really cared about people who commit suicide.

They're selfish bastards.

Geoff Matthews said...

I'm hesitant to accept the regression analysis on behavioral economics. The relationship isn't nearly as neat.
And no confidence intervals? I'd like to know what those are.

Kirk Parker said...

People, this is Russ Douthat. Why are we paying the slightest attention to him?

SomeoneHasToSayIt said...

lol, what a GIGO study!

elkh1 said...

Everyone religious, everyone owns a gun... will become suicide bombers.

Chip S. said...

I estimate the proportion of comments in this thread made by people who actually read the underlying post by Sean Trende to be 5%, w/ a 95% confidence interval of 2.5%.

elkh1 said...

How about no guns, no religions? Like Mao's China?

The state will commit "suicide" for you.

elkh1 said...

EMD said...
"...people who commit suicide are selfish bastards."

Au contraire, they're the least selfish bastards. They stop producing carbon dioxide, doing their parts in saving humanity from global warming; they stop their demands on Obamacare leaving more for their country men. Leftists should love the suicide success.

Original Mike said...

"with some actual regression analysis"

i.e. with a patina of actual science to convince the ill-informed (including themselves) that their results have meaning.

Chip S. said...

a patina of actual science to convince the ill-informed (including themselves) that their results have meaning.

What are your specific criticisms of Trende's post?

Original Mike said...

@Chip:

"in a hypothetical state ..."

I find the concept of putting numbers on a "hypothetical state" to be amusing. Amusing, that is, until somebody tries to drive policy with them. At that point it's no longer funny.

A) I don't believe these can be any more than guesses, and

B) isn't it convenient that no one can test their results?

Chip S. said...

Trende is basically making just about every single point that commenters here are criticizing him for being unaware of.

Since nobody seems to have the energy to click thru to it, here's the gist of Trende's post (he does not present it as a "study", but as a cautionary tale about statistical inference):

• Ross Douthat speculated (@ NYT) that "the recent rise in suicide rates nationally [was a possible consequence of] the retreat of society from traditional institutions, including religion and marriage." IOW, a standard socon argument.

• Nate Cohn (@ TNR) called bullshit on Douthat on the grounds that there is no statistically significant correlation b/w suicide rates and religiosity, marriage rates, or homeownership rates on a state-by-state basis.

• Sean Trende (@ RCP) points out that "once you control for gun ownership, religiosity does have a statistically significant relationship with suicide". IOW, he's pointing out that the claim that religiosity is uncorrelated w/ suicide is nothing more than an example of omitted variable bias. Trende is blaming Cohn for "junk statistics".

Moreover, Trende is careful not to push any social policy: "We must also remember that all we’re really finding here are correlations. Making the leap to causation is just that: a leap."

What more do you want him to say?

I suppose he could have said something like, my purpose here isn’t so much to look at the merits of the various claims and counterclaims -- I have no particular dog in this fight, because I doubt I’m going to add much to a question that has been studied and debated extensively by sociologists for decades. Rather, I want to focus on the particular example excerpted above from Cohn’s first piece, because I think there’s something valuable about the use of statistics to be learned here.

Oh wait...that's an exact quote from him.

What seems to have you all worked up is his way of trying to illustrate for dummies how to think about the relative magnitudes of the religiosity and gun-ownership variables. But blame Althouse for front-paging the one quote guaranteed to generate a bunch of ill-informed rants by making it seem like someone, somewhere, wants to use ordinary least squares to take away people's guns.

Original Mike said...

Who's worked up, Chip?

Clyde said...
This comment has been removed by the author.
Rusty said...

Once again Chip S. ruins the grading curve.

Clyde said...

So all religions are alike in terms of their views on suicide? I'd say that's about as true as the assertion that all religions are alike in terms of their views on homicide. Some say "love thy neighbor," some say, "kill thy neighbor if he is an infidel." Their generic "religiosity" is scientifically null and void.

Chip S. said...

Who's worked up, Chip?

Me, I guess.

Original Mike said...

You're right, I didn't delve into it. My comment was just a bigoted one from the viewpoint of a physical scientist whose always been extremely skeptical of the social "sciences". (Oh; I did it again.)

Marshal said...

But blame Althouse for front-paging the one quote guaranteed to generate a bunch of ill-informed rants by making it seem like someone, somewhere, wants to use ordinary least squares to take away people's guns.

I don't think we need to blame Althouse. It's obvious there are people all over this country who justify their desire to take everyone's guns with ordinary least squares or whatever other excuse they think will work.

I think you're interpreting the criticism too specifically at Trende.

CWJ said...

Chip S. @12:28

Normally, I appreciate your comments. And normally I let things slide if a large amount of time has passed, but I will break both those rules now.

Your comment at the above time stamp is completely asinine. I not only clicked through to Trende's post but tried tried to click further on. There is nothing there to rationally evaluate. Just because someone has access to SPSS and can type "regression" doesn't mean that what comes out the other end is worth shit.

The fact that you could type 5% plus or minus 2.5% with a probability attached to it and think you've said something shows that you don't know squat about metrics, statistics, and measurement.

Jeff Teal said...

Now if Trende had just been concise and said that crap statistical analysis of crap hypotheticals produces crap conclusions which drive crap policies I might have agreed that anyone who didn't read the article should have done their homework in Stats.But he didn't so why should I give a crap?

Jeff Teal said...

To put it another way the inferential utility of both the study and the article closely approach zero.But rhetorically the people who don't do math may find either self-confirming.

Jeff Teal said...

Lies,damn lies,and statistics.(Which in this case are just fancy lies)