THE LEARNING OF CHOICES BETWEEN BETS,
Abstract
A choice between bets situation was analyzed in terms of Estes' scanning model. On each trial, the subject was offered a choice between two alternative bets. He was asked to choose the bet which was most likely to win on that trial. Following his choice, feedback was given for both bets. Wins and losses on both bets were programmed independently according to simple non-contingent random schedules. The principal independent variables were the values of the probabilities of wins and losses on the two bets. All pairwise combinations of .8, .5, and .2 were employed, yielding six main experimental conditions. An additional condition was run in which subjects received feedback only for the choice made on each trial. Twenty-four subjects were assigned to each experimental condition. The scanning model adequately predicted final choice proportions for all conditions. A linear learning model in conjunction with the scanning model was applied to the sequential statistics with some success. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 13, 1964
- Accession Number
- AD0430567
Entities
People
- Morton P. Friedman
Organizations
- University of California, Los Angeles