One hundred and forty-seven college students and 50 sixth-grade children, each being classified into three groups according to piano-playing skill, rated goodness of five sets of four performances (original (O), monotonous (M), swapped (S), and artificial (A)). M was constructed by electronically equalizing the intensity of the sounds, S by swapping the intensity among a half of the sounds, and A by assigning the intensity to each sound according to three rules. Immediately after the initial rating, the college students were exposed to the same sets of performances and were asked to rate again and to choose from eight adjectives the one describing best each performance. Analyses of ratings and of adjective selection revealed that even the nonlearners could differentially and appropriately evaluate three performances (O, M, and 5) as to A, the better skilled, the more negative was the subjects' evaluation. Coherence of ratings across the five sets for best performance was greater among the better skilled and the older subjects. Stability of goodness ranking between the two ratings by the college students was higher among the experienced than the nonlearners.
