Choice experiments to improve predictive power for policy makers
Funding: 2010: $50,000
Project Member(s): Street, D.
Funding or Partner Organisation: Australian Research Council (ARC Discovery Projects)
Start year: 2010
Summary: Choice experiments are used extensively by decision makers to predict the potential public benefit of proposed policy changes. However, often the choice experiment used only allows for the estimation of parameters in a main effects model. When this model is inappropriate, typically because of the existence of interaction terms, the parameters are biased and inaccurate predictions can result. The reason that decision makers use these designs is that appropriate designs for the estimation of more complicated models are not available. This project will construct suitable designs for these models, thus leading to more accurate predictions and better decision making.
Keywords: Stated preference choice experiments; Discrete choice experiments; Orthogonal arrays; Balanced incomplete block designs; Health economics;
FOR Codes: Statistics, Pure Mathematics, Applied Economics, Health Policy Economic Outcomes, Expanding Knowledge in the Mathematical Sciences, Public Services Policy Advice and Analysis