Test performance in optional shift and configural acquired equivalence are positively correlated.

In 2 experiments, participants completed 2 computer-based tasks: a configural acquired-equivalence procedure and an optional-shift procedure.Both revealed that test performance was positively correlated, even when controlling for non-specific variables.This finding supports the suggestion that a common mechanism underlies performance in both tasks.Experiment 2 included eye-tracking to the stimuli used in the task.We found that participants who attended to the predictive compound elements in the optional-shift training went on to show stronger attentional-set effects in the subsequent test.The relationship between attention and performance is considered by reference to attentional and non-attentional learning theories.