Tuesday, December 7, 2010

Suicide Statistics in the United States

David Brooks column in the NYT today is devoted to a series of interesting links to social science papers (an unconventional approach to newspaper column writing that I heartily approve of).  Following his links, I came across a fascinating paper Family Stigma, Sexual Selection and the Evolutionary Origins of Severe Depression's Physiological Consequences. The basic argument of the paper is that depression is a mechanism that has evolved in order to cause individuals of lower fitness than their family's generally to remove themselves from the population to avoid adverse sexual selection against their kin due to reputational damage.
It is now known that in addition to their psychological and behavioural symptoms, chronic depressive illnesses have damaging effects on both the immune and cardio- vascular systems. Given that similar patterns are found in other species, there is a need to determine how such seemingly disadvantageous mechanisms could have persisted over evolutionary timescales. This paper offers an answer by applying inclusive fitness theory to the process of mate selection. To mate selectors, the value of direct observation is limited by the phenotype/genotype disparity. Although it does not seem to have been previously suggested, there is evidence that natural selection has partially circumvented this limitation by favouring the use of the stockbreeding technique of obtaining additional genetic information from a candidate’s close kin. One implication of this is that individuals much less well fitted than their kin can have a negative impact upon the reproductive success of their abler relatives. Indeed, the point could be reached at which the aggregate reputational damage they inadvertently cause imposes reproductive penalties across their kin group substantially in excess of their own likely reproductive contribution. Under such circumstances, as a form of damage control, inclusive fitness considerations would favour the emergence of autonomic mechanisms serving rapidly to eliminate the less well fitted individual. Severe depression, given how life-threatening its symptoms would have been in any imaginable Era of Evolutionary Adaption, could have evolved to bring this about. In this paper, literature relating to: mate choice among peacocks, an island population affected by a debilitating genetic disease, the stigmatisation of families of psychiatric patients, the determinants of self-esteem, and the psychological factors underpinning human health and happiness has been reviewed against four hypotheses derived what the author terms "stigma theory". Although much remains to be done, the initial findings are strongly supportive.
This is a pretty intriguing idea.  In general, it seems that depression and suicide are potential indicators of how stressed a population is - when it gets harder to "make it" in some acceptable way, one might hypothesize that the rates of depression and suicide go up.  Thus, in turn, as peak oil, climate change, financial crises, and other societal challenges bite harder one might expect that to be reflected in suicide rates.  Since depression diagnoses are subject to changes in diagnosis, I started by looking at suicide stats (which seem likely to have been fairly accurate for a long time).  I found data from the Centers for Disease Control, where I chose to look just at the mid adulthood rates (ages 25-54), and the Statistical Abstract of the United States (which has data for the entire population).

Here are the (age-adjusted) rates:

This data would seem to tentatively bear an economic explanation: following the crises of the 1970s (when suicide rates went up), there was an era of relative prosperity in the 1980s and 1990s, culminating in the Internet boom of the late 1990s (when suicide rates bottomed out).  Then in the 2000s, a recession was followed by stagnating incomes and then an energy crisis.  Suicide rates started up again, particularly among adults.

The data end in 2007, so it will be interesting to see if they went up again in the great recession.

There are of course non-economic factors going into this series.  For example, new generations of antidepressants became widespread in the 1980s and 1990s, and one might imagine (or at least hope) they had some effect on preventing suicide.

It turns out there are large regional patterns in suicide:

I have to say it's not at all obvious to me what would cause that particular regional pattern.


brett said...

"I have to say it's not at all obvious to me what would cause that particular regional pattern. "

states with large cities:low
relatively unpopulated states:high

Stuart Staniford said...


Yes, that does look plausible on an eyeball. I have those stats somewhere, so I can play with it further.

John Callender said...

I wonder if there might be a correlation between the states with high suicide rates and the states with high rates of gun ownership. So one could posit that the people in the mountain West are not particularly more prone to suicidal thoughts than coastal dwellers, but that they have the means to act on it successfully when they have them.

That's just the first thing that occurs to me, though there are also a lot of other possibilities. The suicide rats do seem to hew pretty closely to population density (denser population = lower suicide rate), as Brett points out. It would be interesting to break the numbers down by some other factors as well (socio-economic, race, education level) and see how those correlate or fail to do so with suicide rate.

Thanks for the interesting item.

latent sanity said...


I'd love to see more recent numbers ... although I suspect that along with the doublespeak of recovery, we will be told the doublespeak of a significant uptick in accidents ...

On the ground, as one who has had friends and acquaintances diagnosed with depression, I am well aware of the widely-recognized side effect of antidepressant drugs being a loss of libido ... the patients themselves often wonder how this is supposed to make them feel better. But an honest appraisal of psychiatry's true purpose does not include "curing" anybody, does it?

bmerson said...

The correlation is not total, but many of the low-rate states are socially liberal and have significant social safety nets. Many of the high-rate states are, how do you put it, "tough love" states that do not believe in strong social safety nets. It may be that when people are down, if they feel like they have no other place to go for help, or if they feel like asking for help is a socially stigmatic act, then may be more likely to fall back on suicide. Just conjecture.

Landala IF said...
This comment has been removed by the author.
Per B said...

I wonder why white males are by far the most suicidal group.
This goes against the popular idea of white male being the most privileged group in society.
I've heard the argument being made that males are more likely to commit violent act, but there are other ethnic groups more likely to be involved in violent acts with much lower suicide rates.

I'd love to hear any and all theories on this.

James said...

RE: the Rocky Mountains, there was a recent study that demonstrated a positive correlation between living at altitude (less oxygen) and suicide rates. This was a connection that no one seemed to have investigated before. Simple Google search should turn it up.

Dale Goddard said...

Here's my interpretation of the map.

The larger the population, the less likely prospective mates are to perform familial observation. Thus in larger populations, unfit members less frequently experience the effect they may have on gene-mates' reproductive prospects.

Said differently, in smaller communities one is more likely to be aware of, and make family-based judgments about a prospective mate, and conversely to be aware of them being made about one's own contribution to family fitness. By contrast, in larger communities one can disappear into the crowd to effectively disown an "under-performing" sibling, removing the social pressures that this model suggests would otherwise act to induce the under-performer's depression.