How would you feel if your art quilt was judged via a simple questionnaire rather than a panel of experts?
I read Meehl’s famous book Clinical vs statistical Prediction: A theoretical analysis and a review of the evidence many years ago. I was reminded of it recently by a discussion in Kahneman’s fascinating book, Thinking, Fast and Slow, in itself a further treatise on the parlous and impossibly illogical state of human decision making!!
Meehl showed with numerous examples that in many fields a simple algorithm could make a better predictions than could experts in that particular field. This included doctors re prognosis (remember the Apgar score they use to evaluate newborns? It has saved many infant lives), wine-growers predicting how good a wine a particular crop will make, stock-brokers (yes! Wall street could give up tomorrow and computers calculate the best buys and sells and do it better!), financial analysts, sporting events, recidivism rates etc etc. When I think of how much money we spend on these “fortune tellers”, instead of on solid research and development into clean energy and so on, my mind doesn’t cogitate, it boggles!
So I started wondering if an algorithm could be developed for judging art, specifically a quilt show…or even if that would be a good thing? We have all known of amazing quilts that weren’t accepted to shows where they should have been – and duds that were included to everyone’s disgust (except I presume to that of the maker!!) Would they have got in if they were assessed simply by a 6 step questionnaire? It would also be a useful way of assessing one’s own work – which babies need help and which will be stars? I know I’m not alone in wondering which of my art works is the strongest.
Meehl concluded from his meta analyses that in order to achieve the best predictions, decisions should be based on formulae, especially in low-validity environments (like an art show). What we also know is that the algorithm doesn’t have to include complex weighting – it doesn’t make any difference according to Dawes’ article “The Robust Beauty of Improper Linear models in decision making”.
Weighted complex combinations are no more reliable than simple ones.
Of course “experts” are extremely hostile to these ideas, they don’t like to think that all their expertise and judgment and sensitivity counts for very little. And they are skilled in limited, local short term situations, but longer term predictions are better assessed by a mechanical combination of a few variables. However, many have so much invested in their expertise that it makes it very difficult for them to accept their weaknesses as well as their strengths.
Okay – so which variables would we pick for judging a quilt? Six is enough. They should, if possible, address different aspects of the work so that there is not too much overlap. Once the six dimensions have been chosen then a couple of questions for each one could be formulated. For example, for me one of the important things is whether or not the piece can hold my interest – so the questions might be:
1a. How long did I look at this quilt when I first saw it?
1b. Did I come back to look at it again?
A second variable I think important would be something I’d call “freshness”. Questions might be:
2a. Have I seen something like this before?
And so on….so let’s see how much consensus as to important variables we would have. So please send in your ideas!! What characteristics of a quilt, or any work of art actually, are the most important? Let’s see if between us we can devise the Ultimate Quilt Judging Algorithm!
And, if you have been, thanks for reading!! Elizabeth