RESULTS
The current empirical research tries to gain insight into the heuristic called
“representativeness” by Kahneman and Tversky (1972) (henceforth abbreviated K&T). K&T claim that, according to this heuristic, the subjective probability of an event, or a sample, is determined by the degree to which it: (i) is similar in essential characteristics to its parent population; and (ii) reflects the salient features of the process by which it is generated.
A standard question, studied over and over again in numerous research papers, deals with sequences of binary outcomes, such as boys and girls (K&T), heads and tails (Hirsch & O’Donnell, 2001), and so on. Classical mistakes made by students are related a.o. to the “ordering fallacy”. K&T claim that this mistake occurs because students want the sample to be similar to the parent population. But what is the parent population of an ordered outcome of six tosses of a coin? Can this outcome be described as a simple random sample with replacement from a box containing two identical tickets, one with H written on it and one with T? If a student is reasoning along these lines, then, wrongly identifying HTHTTH as having higher probability than HHHHHH could be attributed to a misconception of representativeness. However, the student could be convinced that the parent population for samples yielding ordered n-tuples is itself an (ordered) n-dimensional random vector. For our example, this could be conceived as randomly drawing from a box containing 26 identical tickets. Why the HTHTTH ticket then should have a higher probability of being drawn than the HHHHHH ticket is not clear at all, but it certainly is not easily explained by a misconception of representativeness. It might be a misconception about the parent population.
In trying to discover students’ reasoning and (possible) misconceptions of representativeness, we first addressed a group of well-trained university graduates (group A). They were given a questionnaire containing only item 1. One week later, they were asked to answer item 3. The results were quite amazing. As could be assumed, those students were aware of the “order fallacy” of item 1, having studied probability and statistics at university level. Indeed, the majority (11 out of the 14) correctly answered the question, in which is stated explicitly that order is important. Item 3 however is different in the sense that no mention is made of any ordering in the outcome. Indeed, it only asks for outcomes “like” some given result, giving the student plenty of opportunity for considering all “similar patterns”. A large group of students however completely missed this point, and stuck to the answer they had given on item 1 the week before (see Table 1.)
Further investigation was undertaken with the students of group B. They were asked to answer two questionnaires, at the same occasion. The first questionnaire was the same as for group A (item 1). After having handed it in, they immediately received a second questionnaire, with item 2 and item 3 on the same page, and in that order. The intention was to make the