The statistical power of individual-level risk preference estimation
- PDF / 2,915,882 Bytes
- 21 Pages / 439.37 x 666.142 pts Page_size
- 101 Downloads / 168 Views
The statistical power of individual‑level risk preference estimation Brian Albert Monroe1 Received: 1 March 2019 / Revised: 17 October 2020 / Accepted: 2 November 2020 / Published online: 22 November 2020 © Economic Science Association 2020
Abstract Accurately estimating risk preferences is of critical importance when evaluating data from many economic experiments or strategic interactions. I use a simulation model to conduct power analyses over two lottery batteries designed to classify individual subjects as being best explained by one of a number of alternative specifications of risk preference models. I propose a case in which there are only two possible alternatives for classification and find that the statistical methods used to classify subjects result in type I and type II errors at rates far beyond traditionally acceptable levels. These results suggest that subjects in experiments must make significantly more choices, or that traditional lottery pair batteries need to be substantially redesigned to make accurate inferences about the risk preference models that characterize a subject’s choices. Keywords Power analysis · Risk preferences · Experimental economics · Expected utility theory · Rank dependent utility JEL Classification C12 · C13 · C18 · C52 · C90
1 Introduction In response to growing evidence that some subjects in economic experiments violate one or more axioms of expected utility theory (EUT), several alternative models were proposed which allow for the apparent violations. Prospect theory (Kahneman Special thanks to Glenn Harrison, Don Ross, and Andre Hofmeyr for providing comments and feedback on this paper. Electronic supplementary material The online version of this article (https://doi.org/10.1007/s4088 1-020-00098-x) contains supplementary material, which is available to authorized users. * Brian Albert Monroe [email protected] 1
School of Philosophy, University College Dublin, Dublin, Ireland
1Vol:.(1234567890) 3
The statistical power of individual-level risk preference…
169
and Tversky 1979), rank-dependent utility (RDU) (Quiggin 1982), and regret theory (Bell 1982; Loomes and Sugden 1982) are among the best known of these alternative models. Many of the newly proposed theoretical explanations of the apparent violations of EUT have been tested experimentally. A well-known example is the experiment of Hey and Orme (HO) (1994) to test if any of a variety of generalizations (and one restriction) of EUT can explain experimentally collected data significantly better than EUT. HO picked “winning” model specifications for each of their subjects on the basis of the estimates of each model and whether each model can be statistically distinguished from EUT. HO (p. 1322) conclude, “our study indicates that behavior can be reasonably well modeled (to what might be termed a ‘reasonable approximation’) as ‘EU plus noise.’” However, HO (1994, p. 1315) raise concerns that as the number of alternative specifications being tested increases, the probability that EUT will be selec
Data Loading...
