Formation of arbitrary stimulus classes in matching to complex samples: supplementary data.

A follow-up study of work published in 1990 showed that consistent use of Tone-relevant trials during training and testing baselines contributed to the reliable formation of 5-member stimulus classes during matching to sample. With Tone-irrelevant trials during either training or testing 4 of 12 subjects formed such classes, but none did so when such trials were used in both training and testing baselines. This extends our prior work.