The clinical sciences literature is full of examples of an association between two variables being mistaken for causation. How many treatments are based on strong evidence that ‘this’ and ‘that’ are related with NO evidence that ‘this’ causes ‘that’? It is a fundamental error and can lead to ridiculous conclusions. That is not the problem I guess because the apparently ridiculous conclusions are usually rejected. The problem is when a seemingly sensible (well at least to some) conclusion is endorsed on the basis of a (biased) explanation.
To emphasise this issue, and to prepare for the monumental milestone of BiM reaching 1000 Facebook members, Heidi took a big lump of initiative and decided to make a short video. It has been great to watch her swear at her computer, tear her hair out (fortunately she has ample) and huff and puff along the way to producing this. One could suggest that she bit off more than she can chew and then chewed like hell. She has produced a 4 minute movie that is a condensed version of the beginning of a talk I did at the 2010 Noigroup Conference in Nottingham. I was fortunate enough to provide entertainment at what must go down as The Best Conference Dinner EVER. One aspect of the ‘entertainment’ was to do an experiment. We had about 180 volunteers. We collected the data, analysed it after the dinner, prepared some slides and presented it in the plenary session the next morning. 12 hours. World record for sure. Anyway, watch the video to find out some of the intriguing results. Oh, and you could let Heidi know if it is worth trying this again on something a bit less frivolous.