Memory search instead of template matching? Representation-guided inference in same-different performance.

Data from two experiments using transformationally related patterns forming rotation/reflection equivalence sets of different size's are presented. In a same-different experiment, two successively presented patterns were judged as same when they belonged to the same equivalence set, and different otherwise. Reaction times depended on the size of the equivalence set to which the patterns belong. This was the case, even when they were physically identical. Results are interpreted within a framework of memory-guided inference which assumes that participants perform a memory search within the equivalence sets activated by the compared patterns instead of comparing these directly. In Experiment 2, goodness ratings were required which are shown to reflect a second component of information processing.