The effect of stimulus compounding in paired-associate learning

Abstract

Summary This experiment was undertaken to determine whether or not empirical summation or compounding would occur in paired-associate verbal learning. To this end, seven groups of 22 S s were run on a nine-unit paired-associate list in which the stimuli consisted of colors, nonsense syllables, or both, and the responses were high-frequency words. The two experimental groups were given five trials on nonsense syllables (colors), then five trials on colors (nonsense syllables), and then were tested to the combined stimuli (nonsense syllables + colors). Two control groups were trained in the same fashion and were then tested to the most recent single stimulus. Two other control groups were given ten trials to nonsense syllables or colors alone and were then tested to their appropriate single stimulus, while a third group was given ten trials to the compound stimulus and was tested to that compound. The results showed that a definite advantage was gained by the groups that received the compound on the final test after having been trained to each of its elements. This was evidenced by the fact that the experimental groups did not show the decrement from the end of training to the final test that all other groups showed, and by the fact that they differed significantly from the control groups that received comparable training but were tested to a single stimulus. The results also showed that this advantage was almost as great as that gained through receiving a single stimulus for ten trials, and, finally, that learning is facilitated by the presence of two cues in the stimulus as opposed to one cue. A pattern analysis of the data indicated that the concept of summation could not account for the advantage gained by the experimental groups, and that the most likely alternative explanation is that the advantage is gained simply by having more than one cue available to recall the response. Finally, the facilitation of learning by the presence of more than one cue in the stimulus was discussed. It was suggested that the superior performance of the dual-stimulus group could be accounted for in a similar manner, namely, that the S selects one or the other stimulus dimension for a particular response and thus produces his own functional stimulus.