STRATEGIES IN CONCEPT IDENTIFICATION STOCHASTIC MODEL AND COMPUTER SIMULATION I

TAROW INDOW; SAYOKO SUZUKI

doi:10.4992/psycholres1954.14.168

STRATEGIES IN CONCEPT IDENTIFICATION

STOCHASTIC MODEL AND COMPUTER SIMULATION I

TAROW INDOW, SAYOKO SUZUKI

Author information

JOURNAL FREE ACCESS

1972 Volume 14 Issue 4 Pages 168-175

DOI https://doi.org/10.4992/psycholres1954.14.168

Details

Published: 1972 Received: - Available on J-STAGE: February 24, 2009 Accepted: - Advance online publication: - Revised: -
Correction information
Date of correction: February 24, 2009 Reason for correction: - Correction: ABSTRACT Details: Wrong : A total of 455 cases, 5 tasks by 91 subjects, of the simplest form of concept identification with stimuli, each consisting of 5 letters, were stratified into 8 groups according to strategies used by the subjects. Two representative strategies were identified through verbal reports of the subjects and also differences in performance: whole-letter strategy to pursue all of an appropriate set of letters simultaneously and one-letter strategy to test one letter at a time and without replacement. Both strategies were simulated by a computer. All the results, the whole cases, 8 separate groups, computer simulations, were also fitted by the original form of Bower-Trabasso stochastic model (1964). Two main findings were as follows: there are substantial individual differences in regard to strategy and the stochastic model gives apparent good fits to each of the heterogeneous set of data which were generated through qualitatively dif-ferent strategies, provided the value of a parameter, the probability of hitting the correct hypothesis, is appropriately defined.
Right : A total of 455 cases, 5 tasks by 91 subjects, of the simplest form of concept identification with stimuli, each consisting of 5 letters, were stratified into 8 groups according to strategies used by the subjects. Two representative strategies were identified through verbal reports of the subjects and also differences in performance: whole-letter strategy to pursue all of an appropriate set of letters simultaneously and one-letter strategy to test one letter at a time and without replacement. Both strategies were simulated by a computer. All the results, the whole cases, 8 separate groups, computer simulations, were also fitted by the original form of Bower-Trabasso stochastic model (1964). Two main findings were as follows: there are substantial individual differences in regard to strategy and the stochastic model gives apparent good fits to each of the heterogeneous set of data which were generated through qualitatively different strategies, provided the value of a parameter, the probability of hitting the correct hypothesis, is appropriately defined.

Date of correction: February 24, 2009 Reason for correction: - Correction: PDF FILE Details: -

Abstract

A total of 455 cases, 5 tasks by 91 subjects, of the simplest form of concept identification with stimuli, each consisting of 5 letters, were stratified into 8 groups according to strategies used by the subjects. Two representative strategies were identified through verbal reports of the subjects and also differences in performance: whole-letter strategy to pursue all of an appropriate set of letters simultaneously and one-letter strategy to test one letter at a time and without replacement. Both strategies were simulated by a computer. All the results, the whole cases, 8 separate groups, computer simulations, were also fitted by the original form of Bower-Trabasso stochastic model (1964). Two main findings were as follows: there are substantial individual differences in regard to strategy and the stochastic model gives apparent good fits to each of the heterogeneous set of data which were generated through qualitatively different strategies, provided the value of a parameter, the probability of hitting the correct hypothesis, is appropriately defined.

Corresponding author

Correction information

Register with J-STAGE for free!