Executing Complex Cognitive Tasks : Prizes vs. Markets

Execution of complex cognitive tasks is often analyzed as an exercise of information acquisition and belief updating. We challenge this view in the context of a non-incremental task, namely, the knapsack problem. First, we provide a theoretical argument why Bayesian updating makes little sense in this context. Second, we provide experimental evidence against the Bayesian approach by comparing the quality of problem solving under two treatments: prizes; markets. We find that Bayesian theory cannot make sense of the data: both systems work equally well, while trading is abundant in the market setup and prices are informative but noisy. The experimental data provide suggestions for a new theory of discovery of solutions in non-incremental tasks