We test in a laboratory experiment the theoretical prediction that risk attitudes have a surprisingly small role in distorting reports from true belief distributions. We find evidence consistent with theory in our experiment.