Title: Understanding and Correcting Measurement Error in Measures of Distributional Preferences
Abstract: I perform an experiment using multiple elicitations of the same preference measure (the slider measure of social value orientation) and of prosocial behaviour to investigate the significance of measurement error in estimation of individual distributional preferences. I apply three methods of correcting for measurement error (regression calibration, simulation extrapolation, obviously related instrumental variables) to answer the question: how much does correcting for measurement error increase estimated effects of social preferences on prosocial behaviour?
Sir Clive Granger Building海角黑料University Park Nottingham, NG7 2RD
telephone: +44 (0)115 951 5458 Enquiries: jose.guinotsaporta@nottingham.ac.ukExperiments: cedex@nottingham.ac.uk