The influence of instruction in a two-choice probabilistic learning task under partial reinforcement

Response preferences: A review of some relevant literature.

“…This tendency has been widely reported throughout the experimental literature (Howarth & Bulmer, 1956;Senders, 1953;Senders & Sowards, 1952), but there is considerable evidence that negative recency effects occur just as often (Brown & Overall, 1959;Fernberger, 1930;Grant et al, 1951;Jarvik, 1951). Van Der Meer (1960) proposed that the differences in asymptotic values found between the two conditions of instructions were mainly determined by differences in motivation which made the utility of a correct prediction greater in the case of instructions implying that the process was stationary. Similar results were found by Hyman and Jenkin (1956).…”

“…Furthermore, subjects were indifferent to eventual outcomes because of the simplicity of the probability-learning situation, and "reactive inhibition" may have reduced the frequency with which subjects responded to the most frequent event. Van Der Meer (1960) argued that instructions suggesting that the sequence of events to be predicted was nonstationary would produce a higher asymptotic value of the predictions of the more frequent event than instructions that the process was stationary. He tested his argument in a two-choice situation and found that it was substantially correct.…”

The a priori estimation of subjective uncertainty in human sequential predictive behavior.

“…An same event is lower. To explain this disextensive and systematic overview of re-crepancy, several authors (Goodnow, 1955; search of this kind referring to human pro- Van der Meer, 1960; McCracken, Osterhout bability learning and decision making be-& Voss, 1962;Edwards, 1962; Lee & Janke, haviour has been carried out by Lee (1971). 1965) analysed the hypothesis that Ss con-One important finding of these expert-sider the task not as "gambling situations" ments was the appearance of probability but more as "problem solving situations", matching behaviour under almost every con-i.e.…”