Effects of familiarization, associative reaction time, and meaningfulness of response terms in forward and backward paired-associate learning

Effects of familiarization, associative reaction time, and meaningfulness of response terms in forward and backward paired-associate learning RONALD LEY* and DA VID LOCASCIO State University of New York at Albany, Albany, N.Y. 12203 The purpose of the present study was to determine the effect of familiarization of response terms on the forward and backward acquisition rate of a list of PAs, in which the . response tenns were rated on associative reaction time (RT) and meaningfulness (associative frequency-Mn). Thirty-two undergraduate Ss individually received 20 familiarization trials in which, for 16 of the Ss, the eight CVCVC units of the list were the response tenns of a PA list which was subsequently learned. For the other 16 Ss, conditions were the same, except that the CVCVCs of the familiarization list were different from the units of the PA list. Both RT and Mn were significant response-term variables in the forward anticipation learning task, but Mn interacted with familiarization, while RT did not, i.e., familiarization increased the learning rate oflow-Mn pairs but not of high-Mn pairs. In the backward learning task, high-Mn pairs were learned in fewer trials than low-Mn pairs, but neither RT nor familiarization had an effect. The findings of recent studies of the effect of associative re action time (RT) of response terms in paired-associate learning (PAL) have consistently demonstrated that within levels of equated meaningfulness (M), short-latency RT response tenns are learned in fewer trials than long-Iatency RT tenns (Ley, 1968; Ley & Anderson, 1969; Ley & Locascio, 1970). With respect to stimulus terms, however, the results of Ley & Locascio (1969) indicate that within levels of M, RT of stimulus terms has no effect on the forward anticipation learning of PAs, whereas M has been shown to be a significant stimulus-tenn variable as weil as a significant response-term variable. In view of the high correlation (r =-.80) between Mn (associative frequency) and RT (Ley & Locascio, 1970), the singular finding is that within the same list of PAs, short-Iatency CVCVCs, as response terms, were learned in fewer trials than long-latency terms of equated M; but as stimulus terms, these same CVCVCs had no effect on acquisition rate. The reliability of this finding was supported by a replication of the experiment (Ley & Locascio, 1969), the results of which were completely consistent and almost identical to the first experiment, i.e., within levels of M, RT of stimulus terms had no effect on PAL, but M of stimulus terms did. However, a backward recall test following forward anticipation learning in the replication experiment resulted in a significantly greater recall of the short-Iatency stimulus tenns (response terms in the recall test) than of the long-latency tenns. SimiIarly, *This research was supported in part by a S ta te University of New York Research Foundation faculty research fellowship to the first author. Psychon. Sei., 1970, Vol. 21 (2) high-M stimulus terms were recalled more frequently than low-M tenns. The results of the studies cited are consistent with the general finding that M is a significant stimulus- and response-term variable. Within levels of M, however, it would appear that the effect of RT is limited specifically to response tenns in forward anticipation learning and stimulus terms in backward recall, or, more gene rally perhaps, to that term of the PA which is required to be recalled. If the general case obtains, i.e., if within levels of M, the effect of RT is limited to the tenn to be recalled, it would be expected that the effect of RT of response tenns, following forward anticipation learning, would have no effect on backward recall, whereas M would. The primary purpose of the present study was to test this hypothesis. The secondary purpose of the present study was derived from two somewhal closely related interpretations of the role of M in PAL. Underwood & Schulz (1960) account for the facilitative effect of M through its high correlation with pronunciability, whereas Goss (1963) accounts for the facilitative effect of M through its high correlation with recognition latency. These hypotheses suggest that the more readily a given verbal unit can be pronounced or the faster a unit can be recognized, the sooner the unit can serve as a stimulus tenn and thus facilitate PAL. If M bears the same relationship to response terms as it does to stimulus tenns, as Goss (1963) has suggested, and if, within levels of M, the role of RT is different from that of M, then variables which might conceivably affect pronunciability and/or recognition latency should have a greater effect on PA terms rated on M than on those rated on RT. Although research on the effects of familiarization has not provideu clear and consistent results (Goss & Nodine, 1965), some evidence that familiarization interacts with M of response tenns has been offered by Schulz & M artin (1964). If familiarization faciIitates pronunciation, its effectshould be greater on low-M tenns than high-M tenns, since high-M terms are c10ser to the maximal ease of pronunciability. Similarly, if familiarization decreases recognition latency, its effect should be greater on low-M tenns, since high-M tenns are cIoser to minimal recognition latency. Therefore, if, within levels of M, the role of RT is different from M, the predictect Interaction between familiarization and M would not be expected between familiarization and RT, i.e., familiarization of response tenns should have Iittle or no effect on the forward PAL rate of short-latency (SL) and long-latency (LL) RT response tenns. The secondary purpose of the present study was to test this hypothesis. SUBJECTS The Ss were 32 undergraduate paid volunteers (16 fern ales and 16 males) from the State University of New York at Albany , naive with respect to prior experience in verballearning studies. MATERIALS The materials used in the forward and backward PAL lists and in the practice list were the same as those used by Ley & Locascio (1970), viz, the response tenns of the PA list for the forward task (stimulus terms in the backward task) were eight CVCVC verbal units selected from Taylor's (1959) list: four high-Mn tenns (two with SL RTs and two with LL RTs) and four low-Mn terms (two with SL RTs and two with LL RTs). The stimulus terms on the forward PAL list (response terms on the backward list) were eight two-digit numbers selected on the basis of their approximately equal associative values (Battig & Spera, 1962). Four different forward PAL Iists were constructed, each with an accompanying backward PAL list. These lists all contained the same stimulus and response tenns, but paired in different combinations. Each of these Iists was arranged for presentation on separate paper tapes for use on a memory drum. The arrangement was in four random orders, with the restriction that each of the eight PAs occupy a different position in each order an (...truncated)