Cornell suicides: Noise or something more?

Much has been said on this blog, at The Sun, and elsewhere about the gorge fences greeting us on our return from Spring break, the administration’s most visible response to the rash of suicides this academic year.  The Sun editorialized in favor of the fences, calling them “an unfortunate necessity.”  Dennis basically agreed in this Midweek Post, with Joe and Ollie taking the other side.

I find the fences to be hideous, and my gut reaction is to say they ought never have been put up.  But what is the appropriate response?  The question is not totally subjective.  Let’s assume for sake of argument that the administration can effectively reduce the number of suicides (which is a good thing), but does so at both a monetary and aesthetic cost.  (Otherwise, we should just take maximal suicide prevention measures.)

I can think of a few possibilities:

1.  The so-called “rash” of suicides was just statistical noise.  If people commit suicide independent of whether others do or not, and we assume for simplicity everyone has the same known chance of jumping, then there will be natural fluctuation in the number of suicides per year.  This year was simply an anomaly.  If this is the case, the administration should do nothing at all in terms of policy changes, supposing that they knew the rate of suicides to begin with.

2.  Suicides are correlated, i.e. “suicide contagion” is real.  As long as the effect is fleeting and dissipates after some time, then it makes sense for the admin. to take special action following a suicide.  Whether suicide contagion is empirically observed or not is another question.  (Can anyone point to a good study?)

3.  The administration doesn’t know the rate of suicides.  This is the most interesting case.  Here we are inferring a parameter–the probability that a student commits suicide–from the data–the number of suicides per year.  Trials are no longer independent, since a lot of suicides last year suggests a higher rate, and therefore more of them next year.  This year’s string of deaths has provided additional data about the suicide rate, suggesting that it’s higher than previous thought.  Here too it makes sense for the administration to step up its prevention efforts.

Whether 1, 2, or 3, or some combination applies is up for debate (although we can check 2 empirically).  So is the assumption that the administration can be effective in suicide prevention.  My view is that they most likely can be, and most likely 3 applies.  Yet we can never know.  And even if 3 does apply, anomalies will still occur.  The true, unknown rate could be one per year, and we are now behaving like its six.