The assessment of subjective probability distributions: A training experiment

For this experiment 22 Ss were especially trained to assess subjective probability distributions for proportions concerning student characteristics. Two different assessment techniques were used. There were four experimental sessions. From the second session on, Ss received feedback about their performance in the previous session(s). While Ss' associated by far too much confidence with their assessments at the beginning, they learned to assess more realistic distributions. Furthermore, discrepancies between techniques were reduced.Some considerations for a more effective training procedure are made and the concept of giving feedback only according to the tightness of the assessed distribution is critically discussed.