When it comes to “the inconvenient truth”, actions will speak louder than statistical predictions, says Professor Ian Wilkinson.
Professor Scott Armstrong, a Professor of Marketing at Wharton Business School in the US has challenged Al Gore to a wager of $10,000 that global temperatures will not rise to dangerous levels in the coming decade. The bet was based on a paper he presented at a conference in New York on Forecasting in which he criticises the methodologies used to predict global warming and argues that no change in is the best forecast.
I have long been an admirer and user of Professor Armstrong’s research but here he is in serious error. I’m not sure such forecasts of global temperatures have been made but, as a marketing professor myself, I want to respond on behalf of those who have a better understanding of the systems and methods involved. Basically, he gets things upside down.
He confuses forecasting based on the statistical models of socio-economic systems, with equations describing the way material, chemical, biological, geological and climate systems work which are based on established results in physics, chemistry, biology and geology. Such models are not like those used in market forecasting. People’s behaviour is difficult to forecast because they respond to the results of bets like this and forecasts we make about them, whereas atoms, molecules and creatures don’t read and respond to newspaper reports. Their behaviour is hardwired in.
Climate change models are based on known mechanisms driving the behaviour of our physical and biological world. They are like Einstein’s equation E=MC2, which is not derived from statistical models of the way measures of energy, mass and the speed of light wiggle around together over time and place; it is derived from serious physics theory building on the work of people like Isaac Newton and James Clerk Maxwell and confirmed by carefully carried out experiments.
The same goes for equations describing chemical systems, including carbon, as well as biological, geological and climate systems. The equations describe identities, not statistical models, that capture the way something becomes something else in a material, chemical and ecological system. The “predictions” are no more than working out of the maths, like saying 2+2 predicts 4, which is plain silly! When we burn fossil fuels, we turn one type of thing (oil, gas, wood, coal) into something else and part of that is carbon emitted into the atmosphere.
We don’t know for certain what exactly will happen with climate change and when, but for different reasons than the Professor suggests. First, our measures of the mechanisms involved are not precise and complete. We don’t have weather stations everywhere measuring rain, sunshine and heat, and there is another problem – the systems involved are non-linear and potentially chaotic. You cannot simply add a bit of an effect on the outcome due to this and a bit due to that and derive a result – but this is what statistical models do, by and large. Such models are statistical only in the sense that numbers are involved.
Science is not about taking a vote among scientists about what is right. Einstein didn’t survey physicists to determine if it was E=MC2 or E=MC3. It was established based on evidence and theory and knowledge about how things work. This is what Professor Armstrong fails to grasp along with many others, because the truth is challenging and inconvenient to some. But it is not inconvenient really.
We have the means to do something about it and many are concerned and want to act. By facing up to the challenge, we can unite us against this common foe and ignite innovation and opportunities of all kinds
(Previously published in Australian School of Business Magazine 2006)
