I find it sad when so much is hung on correlation analysis, including, I fear, the times when I have been responsible for this misdemeanour myself!

I recommend reading “Pay No Attention to the Model Behind the Curtain” by Philip Stark, the University of California statistician. This book chapter explodes some of the myths about the validity of statistics.

The great assumption in this tornado study is that proportionate change denotes causality – it doesn’t, or at least it doesn’t prove a direct relationship without the intervention of other controlling variables.

The paradox is that correlation studies do suggest that with earthquakes seismic energy is only a moderately important factor, and corruption tops the list of causative factors (Escaleras et al. 2007, Ambreseys and Bilham 2011, etc.). The important lesson here is not the misleading quod erat demonstrandum of the tornado study, but the invitation to search for the mechanisms that actually prove the case.

Best wishes to all,

David Alexander

PS: In decades of studying the epidemiology of disasters I have generally found that trends and patterns in the USA are not replicated much elsewhere in the world – even in Canada. It would be interesting to establish once and for all why that is so.

From: Radix [mailto:[log in to unmask]] On Behalf Of James Lewis
Sent: 28 May 2017 14:34
To: [log in to unmask]
Subject: Re: Tornado Casualties Depend More on Storm Energy Than Population

On the question of 'the role of society in face to that natural phenomena' and if I were to add my comment also, it would be that 'storm energy' versus 'population' is an over-simplified view of the issue. Of course storm energy would be significant but also so would what people, ie: population, had done or had not done in a tornado-prone environment. In other words, behavioural factors would be part of the issue in this example of 'adjustment' to hazards. Have we not seen that term somewhere; didn't it come from natural hazards research in the USA during the 1960s (culminating in eg: White, Gilbert (Ed) Natural Hazards: Local, National, Global. OUP New York London Toronto. 1974)?

Perhaps after all "There's no new thing under the sun"....

James

James Lewis Datum International www.datum-international.eu