I have heard/seen/read several articles and interviews about the so-called debate between Gerd Gigerenzer and Daniel Kahneman (for example here) about what to do to help people make better decisions. I was so interested that I went out and read Gerd’s new book and re-read Kahneman’s. I lost a few days, but it was worth it. I agree that there are differences, but I don’t think that it is the dichotomy that people are using to drive traffic to their web sites.
A rival psychologist has published a book debunking the behavioural economics of Daniel Kahneman and the men behind Nudge, who, along with the authors of Freakonomics, were once the PM’s pet thinkers. So how do you choose between them?
Here is the dispute in a nutshell. Kahneman is the most famous (Nobel Prizes tend to do that) so I will start there. He models decision making using a hot and fast System 1 that takes shortcuts using heuristics, is biased by emotion, and therefore makes less optimal decisions and a cold and slow System 2 that is more logical and optimal. Therefore, to help people make better decisions, we should push them towards using System 2 or use smart defaults and paternalistic policies to nudge them in that direction.
Gigerenzer finds two main faults to this approach. First, System 1 is often correct when System 2 fails. This happens because System 2 is limited by working memory and many of the complex decisions we make in life go beyond what working memory can handle. So our gut System 1 decisions, although we can’t articulate how we came to them, are sometimes better anyway. Second, he finds that many of the people who use System 1 inappropriately simple lack statistical knowledge and if we taught them some basic techniques, they would be much better. He highlights the difference between risk and uncertainty – which is one of the causes of the financial market crash of 2006-8. He also points to some of the intentional tricks that companies use to sell products that intentionally mislead.
I will share one example because I thought it was really telling (and unfortunately is used in real life).
A theoretical cancer screening test:
- With screening, the 5-year survival rate is 100%
- Without screening, the 5-year survival rate is 0%
Sounds good right? But this is because of poor education on survival rates and a trick of lead time. Imagine you have 100 people, all screened at 60 years old. They all die of cancer at age 70. What is the 5-year survival rate “from” the screening? It is 100% (because they lived more than 5-years from the date of the screening-based diagnosis). Let’s say that without screening we see the symptoms emerge at age 67. What is their 5-year survival rate? 0% (because none of them lived longer than 5 years from the date of the symptom-based diagnosis). But it is the same 100 people. The truth is that none of them got any benefit whatsoever.
But the reason I think this debate is more manufactured than real is that both experts, when push comes to shove, would admit that both sides are correct. The answer is much more nuanced and complex than a single answer could provide. Sometimes, a fast System 1 biases a better System 2 and a paternalistic nudge might be useful. Other times, better use of risk, uncertainty, statistics, and education could be better. Let’s use all of the tools in our toolbox, rather than create false debates.