Dismissing Durkheim . . . and Sociology

May 23, 2013
Posted by Jay Livingston

There it was again, the rejection of sociological thinking – not because it’s wrong or because it offers no effective policies, but just because it doesn’t make people feel better. 

A Newsweek article on suicide makes the obligatory hand-wave:

Sociologists in general believe that when society robs people of self-control, individual dignity, or a connection to something larger than themselves, suicide rates rise. They are all descendants of Emile Durkheim, who helped found the field in the late-19th century, choosing to study suicide so he could prove that “social facts” explain even this “most personal act.”

That’s 58 words in a 6600-word article, not so much a shout-out as a mumble-out.   I exaggerate.  The article does cite sociologists Julie Phillips and Sherry Turkle, and it tosses around statistics about suicide rates by age, sex, race, and birth cohort.  Still, the paragraph that starts with sociology and Durkheim ends with this curt dismissal of sociology because it cannot play to people’s feelings by predicting individual cases:

But when someone’s son dies by suicide and the family cries out for an answer, “social facts” don’t begin to assuage the pain or solve the mystery. When a government health official considers how he might slow down the suicide problem, “society” is a phantom he can’t fight without another kind of theory entirely.       

That other theory, needless to say, is focused on individuals, and the center of the article is a psychologist, Thomas Joiner, whose first job was to identify– and quarantine! – those who would otherwise kill themselves. 

He got to regularly look suicidal people in the eye, only this time he did so knowingly, as a therapist, and with a decision to make: which of these people were risks to themselves? Under Texas law he was allowed to lock people up if they were.

The article gives no data on whether Joiner was actually able to pick out the truly suicidal.  I would guess that he had a Texas-size false-positive problem. That problem comes with trying to predict and change individual behavior.  It is much more socially beneficial and accurate to think in terms of predicting and changing rates of behavior.

If someone’s son dies in a car crash, it might “assuage the pain or solve the mystery” to find out who was to blame – which driver was drunk or momentarily distracted or whatever.  It’s much less comforting to look at aggregate rates.  But when you do, you just might notice that crashes are frequent on this one stretch of road or that crashes are more likely to be fatal in cars without airbags or seatbelts.  Those “social facts” lead to structural policies that can reduce the overall numbers. 

In fact, the absolute number of highway deaths in the US in 2012, despite a 5% increase over 2011, was  lower than at any time since the early 1950s.  The rate per vehicle-mile has fallen by 80% since the 50s and 60s.  Most of that decrease – tens of thousands of saved lives each year – came not from identifying fatality-prone drivers but from changing the structure of roads and cars.  We don’t know which individual lives were saved.  We just know that there were a lot of them.