Optymizm a prawidłowości percepcyjne przy ważeniu prawdopodobieństwa

The main question of the presented experiments has been whether the probability weighting function (Kahneman, Tversky, 1979) and configural weighting incorporated in the CPT model (Tversky & Kahneman, 1992) can account for the overweighting of positive outcomes and underweigthing of negative outcomes, the phenomena first observed in the 50−ties (Irwin, 1953; Marks, 1951; – Marks-Irwin effect) and called unrealistic optimism in the current literature. In Experiment 1, carried out with 133 students, the relations between expectations of success, weights attributed to positive and 0 outcomes and the objective frequency of such outcomes were investigated in a skill situation, i. e. a competition for the best student paper. It was found that: (1) expectations of success depended strongly on the objective frequency of prizes, (2) the best outcomes were weighted in agreement with predictions of the CPT model, but (3) the same was not true for the worst outcome, i. e. 0. The results, however, could not be easily generalised because the situation was constructed such that no negative outcomes were present and 0 outcome was always the most probable. To extend studies to lotteries with negative outcomes and wider range of probabilities an additional experiment was carried out. In Experiment 2 284 students made choices between 10 pairs of 5-outcome mixed lotteries. These lotteries were constructed such the absolute utilities (calculated from CPT) of positive and negative outcomes were equal, and positive and negative outcomes had the same probabilities of occurrence. According to CPT, such lotteries should be always rejected. because negative outcomes are weighted more heavily, yielding negative overall CPT values of the lotteries. The results, however, did not support such predictions – the acceptance rate was close to 1/ . In general, the binominal distribution of acceptance of gambles with negative CPT values might suggest that some people assign optimistic weights to outcomes, i. e. overweight probabilities of positive outcomes and underweight probabilities of negative outcomes. In particular, individual differences in dispositional optimism (measured with the LOT scale) can reinforce such tendency.