Sunday, 27 May 2018

Statistics, damn statistics

I attended a workshop on mental health first aid recently. I'm not good at dealing with intense emotions but I am glad I went. I came in with a certain amount of trepidation but a lot of the stuff that were discussed are things that I've already come across at some point in my career as an instructor.

One thing that left me with some bad taste in my mouth was (and always) is how statistics on suicide are presented: 30 times more than, this group is more likely...etc.

This got me thinking how utterly useless the current statistical approaches are to the actual discourse of suicide prevention. I know a few statisticians and how much they love (actually 'love') their datasets and how intertwined they become to their own analyses rather than the actual subject of study. All good and well, I suppose.

But, none of it is new news. There is an almost complete absence of original insight to be gotten in examining events that have already happened simply because what is likely to happen is never addressed. Useless, as I said.

I think what needs to happen is to examine how many households have had family member(s) commit suicide; what are the income levels of these households; what are the rates of bullying in any given school in relation to communities where suicide rates are high; what levels of follow-up occur when individuals come out of the health and criminal systems; etc. etc.

Rather than demographic cohorts, I think we should expect statistics to look at actionable aspects of these sad numbers (I mean, policy- and program-wise): quality of life, for individuals and communities. I mean, start with things that policy/program planners and local and territorial gov'ts can act upon strategically.