One of the things I have just done is out myself, by linking explicitly to my university profile. The fact I blog as a reformed tory is no great secret. But it means that I can talk a little more about the science part of the title.
Ingvild M Tollefsen, Erlend Hem and Oivind Ekeberg have just systematically reviewed all studies on the reliability of suicide statistics. This is not a trivial question. One of the outcomes that clinicians try to avoid when dealing with those who are mad or in despair is suicide. The rate is much higher in psychiatric patients than the general population. There is a entire literature on the variability of suicude rates by country and over time.
What these authors, claim however, is that the official statistics are unreliable, and tend to under report.
SUmmary table from BMC Psychiatry
One of the basic issues in epidemiology is reliability. Your statistics may not be valid — I’d argue, for instance, that the US race classification is not valid — but they need to be reliable. You can’t have validity without reliability. It looks like the human bias to call suicide anything but (and avoid the shame of this event) is hiding the problem from populations, and not allowing researchers to be certain about anything. Clear, reliable data helps generate hypotheses, whcih can then be tested and out of the detritus of failed explanations we winnow out truth.
But without reliability, the scientific method does not have much power. We are all running blind. We cannot suggest ways to improve what we measure poorly. This paper confronts us all: the institutionalised practice of silence hurts the living, because we cannot learn from the tragedies of the past.
Reliability of measures matters. Methodology matters. And anyone who claims to be working in the social sciences and tells you otherwise is a fool or a liar.
