ACQUISITION AND TRANSFER OF A SECOND-ORDER CONDITIONAL DISCRIMINATION

Abstract

Six types of trials were designad for a second-order matching-to-sampla task with explicit instructions and text options that the subject could use to describe hís/her parformance. On each type of trial, correspondence among tha different elements of tha task (instructions, second-order stimuli, programmad feedback and text options for describing performanca) was manipulated through false or true instructions, feedback, and descriptions. ln tha first experiment, six groups were trained with a single type of trial and two other groups ware axposed te all possible types. Matching performance was affected by false or true feedback, while accurate selection of true descriptions was regulated by falsa or true ínstructions or by the correlation of instructions and correct feedback. Tha second experiment replicatad sorne conditions of the first with observational instead of instrumental training. The effect of feedback preved stronger, thareby weakening the effect of tha instructions en both the verbal description of the trials and, atar, en performance during transfer tests and posttest. Taken togather, the results of tha two expariments do not support the assumption that variables such as instructions, descriptions, and feedback haya specific functions on behavior.