When and why do people avoid unknown probabilities in decisions under uncertainty?

Rode, C., Cosmides, L., Hell, W., & Tooby, J. (1999). When and why do people avoid unknown probabilities in decisions under uncertainty? Testing some predictions from optimal foraging theory. Cognition, 72, 269-304.

Summary

Sets up a foraging ‘system’ to explore decision-making.

In this view, the system is not designed merely to maximize expected utility. It is designed to minimize the probability of an outcome that fails to satisfy one’s need, as per Keynes.

The people who participated in our experiments executed complex decision strategies, ones that take into account three parameters mean, variance, and need level rather than just the single parameter (mean) emphasized by some normative theories. Their intuitions were so on target, that their decisions very closely tracked the actual probabilities of each box satisfying their needs. This was true even though explicitly deriving these probabilities is a nontrivial mathematical calculation.

Comment

This gives a foraging setting in which rather than gathering the most food in the long run, the aim is – firstly – to have enough to survive in the short run, and then to build up a surplus in the long run. It rightly notes that this calls for a different approach. Confusingly (to me) it describes the utility approach as ‘logical’ and ‘mathematical’, from which some seem to infer that trying to maximize sustainability is not.

• A strategy that seeks to maximize expected return / return ‘in the long run’ may not be appropriate when there is short-term jeopardy. (As Keynes’ said, ‘In the long run you are dead.)
• It is not logical or mathematical to use a theory whose assumptions / axioms are known to be false, although (according to some definitions) it may be ‘rational’. If one is not certain that the assumptions / axioms are ‘true’, it is not logical or mathematical to avoid the uncertainty.
• Logic and mathematics such as Keynes‘ can cope with short-term decisions, or situations where a balance is needed between short and long-run issues.
• Logically, typical foraging tasks are best met by a population of foragers with different ‘attitudes to risk’. That is, most foragers may take a short term view but some need to take a long term view (to find new food sources). This relies on sharing when the explorers come back empty-handed.

One should also note that the original paper uses variance in a stereotyped way that is not always appropriate, as emphasised by Taleb., who alos discusses the general problem of ‘resilience to tail risk’.

Dave Marsay About Dave Marsay