Suicide and Well-Being. SOC 101, Week 1

March 3, 2019
Posted by Jay Livingston

I begin the semester with Durkheim’s idea of social facts, and I use his example of suicide rates. The rate may be made up of individual cases, but that rate takes on an existence that seems separate from those cases. It is more a property of the society or the specific group. Here are the numbers of suicides and the rate per 100,000 (age-adjusted) in New Jersey for the last four years (CDC)
2014   786 (8.3)
2015   789 (8.3)
2016   687 (7.2)
2017   795 (8.3)
In three of the four years, the numbers are nearly identical, differing by only 9 suicides in a population of over eight million. So it makes sense to think of the rate as something about the state, not about the individuals that make up that rate. Rates in the other 49 states, though they vary widely from state to state, show the same kind of stability. Each year the state produces roughly the same number of suicides.

In case students had missed the point that it’s not about individuals, I remind them, “The 789 people who killed themselves in 2015 cannot be the same 786 who killed themselves in 2014.” I add, “There aren’t many facts in social science that we’re 100% sure of, but that’s one of them.”

My second point is that while we can use individual facts to explain other individual facts, when we try to explain social facts, those same explanatory individual facts often aren’t much help. For explaining the individual suicide, it makes obvious sense to look at a variable like happiness. I’m willing to assume that people who kill themselves are not as happy as people who don’t. But are people in Greece three times as happy as Americans? 

A headline in the local papers a couple of days ago looped us back to that first week of class.

In fact, New Jersey ranked 31st. The headline is referring to a recent Gallup report (here). Gallup calls its measure “well-being,” not “happiness.”  Whatever. As for the happiest or wellest-being  states? Here’s the map.

The map of well-being looks strikingly similar to the map of suicide that I show students in Week 1. The same states that have a lot of well-being also have a lot of suicide. Here is Gallup’s list of the top ten on well-being. I have added a column to show the ranking and rate for age-adjusted suicide.

(Click for a larger, clearer view.)
All but two of the states highest on well-being are in the top twelve on suicide rates. Only Delaware has a lower-than-average suicide rate.

If happiness doesn’t keep suicide rates low, what does? Durkheim’s answer was “social integration.” Unfortunately, Gallup doesn’t have a variable by that name. But the Well-being index is a score made up of five components: Career, Financial, Physical, Social, and Community. The one that seems closest to Durkheim’s conception of social integration is not Community (“liking where you live, feeling safe and having pride in your community”) but Social (“ having supportive relationships and love in your life”). What the scale-makers call Community does not sound a lot like Gemeinschaft. It's more an individual feeling of pride or safety. It does not require actual involvement with other people. By contrast, Social does seem to be a measure of interpersonal involvement.

Since Social seems much closer to Durkheim’s notion of social integration than does Community. So we shouldn’t be surprised that those high-suicide mountain states also rank high in Community. But mostly they are not among the highest in Social. New Jersey, with its low suicide rate, is low on Community (ranked 40th) but high on Social (9th).

There are many anomalies. Colorado, for example, comes out very well on Social and all the other sub-scales of Well-being, yet its suicide rate is 10th highest (tied with Nevada). New York  ranks in the bottom half on four of the five components, including Social, and in the bottom fifth on three of them (Community, Career, Financial), yet it has the lowest age-adjusted suicide rate among the fifty states.

The Gallup numbers do support the Durkheim explanation — not overwhelmingly, but enough for the first week of class, enough to open the door to social  explanations of what seems like a highly personal decision.


Unknown said...

Huh? Isn't the suicide rate per state in the US strongly correlated with gun ownership?

I'd think you'd be hard put to find a "social" aspect that correlates as strongly with suicide as gun ownership.

Am I giving away the answer?

Andrew Gelman said...

1. Scatterplot of suicide rate vs. avg well-being score?

2. Scatterplot of suicide rate vs. gun ownership rate?

Each scatterplot with 50 points, one for each state, using different colors for the 4 regions of the country.

The 2 scatterplots next to each other.