INTRODUCTION

In a series of short-term memory (STM) experiments Conrad (1963, 1964) has shown that sequences of items which are hard to discriminate in noise are also hard to remember, even though presented visually. Analogous effects of intra-list similarity have also been shown in long-term memory (LTM) where several types of similarity have proved to be relevant including similarity of letter structure (Horowitz, 1961) and of meaning (Underwood and Goad, 1951; Baddeley and Dale, 1966). However, Baddeley and Dale (1966) using paired associate learning failed to show an equivalent effect of semantic similarity in STM and suggested that STM may differ from LTM in relying more on acoustic cues and much less on the meaning of the material to be retained. The present study uses the method of serial recall to explore further the role of similarity in STM.

Experiment I compares the influence of acoustic similarity on ordered STM for word sequences with that of semantic similarity.

EXPERIMENT I

METHOD

Design

A separate group of subjects did each of two conditions, A and B. Both groups attempted to recall 24 sequences of five words. In condition A these comprised 12 drawn from a set of eight acoustically similar words (mad, man, mat, cap, cad, can, cat, cap) and 12 from a control set of acoustically different words of equal Thorndike-Lorge frequency (Thorndike and Lorge, 1944) (cow, day, bar, few, hot, pen, sup, pit). In Condition B, 12 sequences were drawn from a set of eight adjectives with similar meanings (big, long, broad, great, high, tall, large, wide) and 12 from a set of eight semantically different words of equal Thorndike-Lorge frequency (old, deep, foul, late, safe, hot, strong, thin). All sequences were drawn at random with the constraint that no word appear more than once in the same sequence. Similar and different sequences were presented in the same random order in both conditions.