Dan Ariely’s The Upside of Irrationality


Jason Collins


September 23, 2013

I rate Dan Ariely’s The Upside of Irrationality: The Unexpected Benefits of Defying Logic lower than Predictably Irrational. Like Predictably Irrational, The Upside of Irrationality is based largely on Ariely’s own work (a good thing). But where Ariely had 15 years of experiments to call on for his first book, for this one he seems limited to a couple of years of newer experiments and the experiments in the reject pile from the first. We then get a lot of filler where Ariely riffs on the theme of the experiment rather than reporting experimental results. It causes this book to feel lightweight in comparison to the first.

That said, I enjoyed the first couple of chapters about work incentives. Ariely and his colleagues ran experiments where they offered incentives to experimental subjects for the performance of mentally taxing tasks, such as remembering a sequence of numbers or hitting a target. But rather than incentives improving performance, high incentives (in the order of several months pay) caused the participants to choke and perform worse than those who were moderately incentivized. This contrasts with experiments that required purely mechanical activities for bonuses, with larger bonuses generally increasing performance.

Ariely related a story about telling a group of bankers about these experimental results, with the bankers suggesting this incentive problem did not apply to them. In some senses, I agree with the bankers, but likely for different reasons. The crumbling in performance witnessed by Ariely and his colleagues was for short-term mentally taxing tasks. In contrast, most bankers, consultants, lawyers and the like are receiving bonuses for a year of effort, making the bonus less salient at any time. But more importantly, the bonuses are heavily tied to the mechanical part of the job - putting in or billing a massive number of hours. I expect there are not many split second decisions that are required to be made with the bonus in mind, and Ariely is overestimating the short-term creativity required.

I find these experimental outcomes somewhat perplexing from an evolutionary perspective. What is the benefit to choking when the stakes are high? One explanation might be that in the environment of evolutionary adaptedness, high-stakes games often ended in death, with choking a signal for the person to get out of there. Another might be that most high-stakes events in that environment simply needed a fight or flight response, not the maintenance of mental coordination.

It was when Ariely got into areas such as dating that the experiments seemed thinner and Ariely was forced to fill more space with his personal views on the subject. I don’t mind a bit of speculation, but Ariely spent a lot of time extending his discussion beyond the experimental context than in Predictably Irrational.

For example, Ariely reported the results of an experiment where, before a speed dating event, participants engaged in a virtual online date where they explored a virtual space together. Those who had earlier participated in the virtual date with their later speed dating partner liked them more. Ariely suggests this indicates a flaw in speed dating setups. But what is the objective of speed dating? Does the increased probability of liking someone due to an earlier virtual date, even though the characteristics of that person have not changed, lead to achievement of the goal of a long-term partner? Or is that familiarity leading them to ignore more suitable people in the room during the speed dating?

Ariely also looked at online dating, and noted that huge amounts of time are expended online relative to the time spent on dates. People do not rate the experience as enjoyable. He saw this as a general indication of the failure of the dating market. I won’t claim that dating markets are perfectly efficient, but again, what is the objective of online dating? I expect it is not enjoyment. If the purpose of online dating is to create a large pool from which the dross can be weeded out, the counterfactual for comparison is going on dates from a smaller pool without that filtering mechanism. Which is the better option? I don’t know the answer to these questions, but I’m not convinced Ariely did either. But as there seemed to be a need for filler around the experimental results, we got a lot of Ariely’s thoughts on these subjects.

I absorbed The Upside of Irrationality in the right way - through an audiobook on my way to and from work. It’s an easy read/listen, has some interesting ideas (particularly early in the book) but seems light compared to Predictably Irrational. I’ll keep Ariely’s next book, The Honest Truth About Dishonesty: How We Lie to Everyone–Especially Ourselves, on my reading list, but I hope it is more dense with research results than riffs on the theme than is the case for The Upside of Irrationality.