Fast and Simple Adaptive Elicitations : Experimental Test for Probability Weighting

We propose a new adaptive procedure for the measurement of decision models. Our procedure bounds the function of the decision model, sequentially halves the maximal distance between its bounds, and incrementally restricts the feasible space of associated parameters. It relies on approximating splines that improve tractability and descriptiveness, enriching results. It grants concise unsupervised experiments: ten choices on average suffice.We apply and validate our procedure in an experimental elicitation of the probability weighting function. Our procedure faithfully captures empirical regularities in probability weighting. In addition, it reveals systematic distortions introduced by standard parametric forms, and sheds new light on the pervasiveness of possibility and certainty effects in close proximity to the probability boundaries (i.e. p=0 and p=1)