REINFORCEMENT-TEST SEQUENCES IN PAIRED-ASSOCIATE LEARNING.
Abstract
Learning of paired-associate items was studied in relation to different repetitive sequences of reinforced (R) trials and test (T) trials. One purpose was to obtain evidence as to whether either learning or forgetting occurs on unreinforced T trials; a second was to adduce principles bearing on the problem of optimal programming of R and T trials. The four training conditions were: (1) R T R T ...; (2) R R T R R T ...; (3) R T T R T T ...; (4) R R T T R R T T ... . Five items were assigned to each condition and the sequences were repeated till a criterion of learning was reached. Two groups of 50 subjects were run; one with nonsense syllable-number pairs and one with nonsense syllable-word pairs. Performance on tests given successively without intervening reinforcement showed no significant change in correct response probability--suggesting that neither learning nor forgetting occurred on T trials per se. The course of learning was, however, affected to a major extent by the ratio of Ts to Rs and by their arrangement in the various repetitive sequences. Learning curves plotted in terms of error proportion on the first T following the n(th) R trial lined up in the order: Condition 3 (lowest), 1, 4, 2. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1965
- Accession Number
- AD0622248
Entities
People
- Chizuko Izawa
- W. K. Estes
Organizations
- Stanford University