No abstract

No abstract

Retarded 5s learned redundant-stimulus paired associates (PAs) in which the stimuli were trigrams and the responses single digits. Each S was responsepretrained and then randomly assigned to one of seven different groups. Each group experienced a different level of PA learning: 1, 2, 3, 5, 7, 2 perfect, or 8 overlearning trials. The 5s were tested for digit recall to each letter from the trigrams presented singly and performance was scored according to either a single-or double-criterion method. Both of these modes of analysis showed first letters to be better recall cues than the other two (after 2 perfect trials) and a large increase in the effectiveness of third letters after 8 overlearning trials. There was, however, one major difference: single-criterion scoring indicated that the amount of single-letter selection increased across trials, but the double-criterion scoring revealed that it was actually two-letter usages which were more frequently employed with increasing degrees of PA learning.

Stimulus-selection findings given consonant-triad: digit paired associates (PAs) were compared with stimulus-selection findings given word-triad :digit PAs. No differences were observed in the specific ways in which 5s tended to process these different kinds of redundant cues: (a) Both sets of triads were processed predominantly by single-cue solutions; and (b) both sets of triads were solved predominantly by selection of first-position or third-position components. Both sets of triads also led to comparable estimates of cueselection efficiency, but there was some suggestion, however, that the word triads led to greater cue-selection consistency (p < .10). It was concluded that, for most purposes, the consonant-triad and word-triad paradigms can be considered roughly equivalent research models.

No abstract