I recently wrote a post called "The Mathematics of Misery," which explored Becker and Posner's idea of effective suicidality. In brief, it may be that miserable people value their lives so little that they rationally accept gambles that happy people would never accept, because they plan to commit suicide if the gamble does not pay off. The authors connect their model to many stories we are familiar with, and find greater explanatory power in their model than in others. The model is particularly appealing to me, because it accords well with my introspective experience as a bona fide, not-breathing-for-a-while-there suicide attempter.
But there are alternative explanations for such Hail-Mary, seemingly irrational choices (not the least of which is actual irrationality). How might we begin to test such a theory?
One obvious choice would be to study the connection between "losing" at a prospective suicide-gamble activity and actual suicides. If unsuccessful participants in the activity do commit suicide at a high rate, the theory is supported - so long as we assume that some higher-than-baseline percentage of the effectively suicidal "follow through" on their suicide plans. (A plan need not be followed for it to be a cognitive reality upon which decisions are based.)
How about investigating whether the strong correlates of suicide are also strong correlates of the candidate "gamble" behaviors?
"Effectively suicidal" behavior may have a different etiology than actual suicide attempts, but there are many reasons to believe that if effective suicidality exists, it is another aspect of the suicide drive, and that they share a common cause. In addition to failed social belonging and burdensomeness, the strongest predictor of suicide in Thomas Joiner's model is competence: the ability to carry out the act of suicide. Competence, as Joiner outlines in his book, is learned; "provocative" preparatory behaviors systematically precede suicides, as if one must train oneself to do it, bit by bit. (This is one function of intentionally cutting one's skin.)
Given that people must achieve competence over time in order to successfully commit suicide, we would expect a continuum of provocative, suicide-preparatory behaviors - and both from introspection and from examples in the Becker-Posner paper, many dangerous gambles are also suicide-preparing.
Let's say we could agree on some strong correlates of suicide, specifically things that seem to cause suicide that also would be expected to cause people to become effectively suicidal (that would make circumstances unacceptable to them, as JasonSL puts it), hence to take a bad gamble/engage in provocative/preparatory behaviors. Let's say I have five or ten candidate behaviors for effective suicidality, and some of them really strongly correlated with the things that suicide really strongly correlates with. Would that be evidence of anything?
Any other suggestions for (a) how to support or knock down effective suicidality (revealed unacceptability) and (b) candidate behaviors?