A typical question therefore reads:Ĭompared with the average student of your age and sex, how likely do you think you are to contract heart disease? In a typical study, participants are presented with a variety of future life events, and asked to estimate their chance of experiencing each event, relative to the average person. Participants report their answer by circling a number between -3 (much less likely than the average person) and +3 (much more likely than the average person). The logic of the test is that, although each participant’s own risk can be greater or less than the average person’s, the average of all participants’ risks should, by definition, be the average risk. Therefore, if the average response on this scale differs from zero, this is taken as evidence for a systematic underlying bias at the group level. The typical result is that, for negative events, the average score is less than zero. This is taken as evidence of optimism, since we desire not to experience negative events.Īlthough the logic underlying the test is sound, in practice its data are compromised by statistical artifacts. Harris and Hahn demonstrated how seemingly optimistic results could be obtained even from agents who had perfect knowledge about their future, through the mechanisms of scale attenuation and minority undersampling. Moreover, for non-omniscient, but non-optimistic rational agents, base rate regression was another statistical mechanism leading to seemingly biased responses.
The detail underlying these mechanisms is provided in, but here we provide a brief description of these mechanisms.
We then go on to conduct three empirical tests to determine what evidence for comparative optimism is observed when controlling for these statistical confounds. The most popular scale used in the comparative method is -3 to +3 (e.g., ).
0 kommentar(er)
