John P. Houston

Verbal R-S strength following S-R extinction

Retroactive inhibition (RI) of R-S associations was investigated by comparing B-A recall following A-B, A-C learning with B-A recall following A-B learning and an interpolated rest activity. Highly integrated and available sequences of A units were employed to avoid greater A unit availability in the experimental group at the time of recall. The results indicated significant RI of R-S associations following A-B, A-C learning

Rehearsal strategies and the primacy effect in serial learning

This study evaluated the effects of three different types of overt rehearsal upon the serial position curve after a single serial learning trial. Ss who rehearsed from the beginning of the list displayed a distinct primacy effect. Rehearsal of successive pairs of items and rehearsal of single items within the list produced minimal and intermediate degrees of primacy, respectively. The curve of a control group, which received no rehearsal instructions, did not differ from that of the group rehearsing from the beginning of the list.

Extinction, new learning and the unlearning phenomenon

Responses in an A-B, A-C retroactive inhibition situation were either from the same or from different classes of materials. The second task involved either the usual anticipation technique or a guessing method in which the S did not experience consistent pairings of particular stimuli and responses. Significant unlearning appeared only when the second task involved the learning of similar materials. While the experiment is not conclusive it does suggest that both the unreinforced evocation of first-list responses and new learning may be necessary for the unlearning effect to appear.

Transfer in repaired S1-R2 and S2-R1 paradigms

Four two-list paired-associate transfer paradigms (A-B, B-Ar; A-B, B′-Ar; A-B, B-Ar′; A-B, B-C) were compared with an A-B, C-D control condition. In accordance with predictions made from a transfer surface described by Houston (1965), negative transfer appeared in all four experimental conditions. In contrast to the surface predictions that maximum negative transfer would occur in the A-B, B-Ar situation, that minimum negative transfer would appear in the A-B, B-C paradigm, and that the remaining conditions would display intermediate amounts of negative transfer, the four experimental paradigms did not differ in terms of second-list learning.

A repaired S1-R2 and S2-R1 transfer surface

A surface was described which summarized the transfer effects expected in two-list paradigms varying in S1-R2 and S2-R1 similarity when the first-list units were repaired in the second list. According to the surface, negative transfer should increase as S2-R1 similarity increases. This increase should be greater when S1-R2 similarity is high than when it is low.

Paced vs. unpaced paired-associate recall

In an attempt to eliminate a potentially confounding factor in the comparison of paced and unpaced recall procedures, Ss first learned either an A-B or successive A-B, A-C paired-associate lists. One week later, Ss were asked to recall the first-list responses. Recall was either paced or unpaced. In contrast to the typical experiment, in which Ss in unpaced conditions are asked to recall both B and C responses, Ss in this experiment were informed that recall of the second-list responses was not necessary. Contrary to the prediction based upon the two-factor theory of forgetting, which states that more retroactive inhibition should be obtained with paced recall procedures, paced and unpaced conditions yielded equivalent amounts of retroactive inhibition. It was suggested that these data may be taken as further evidence for the suggestion that competition at the time of recall is not an important determinant of interference effects.

Retroactive inhibition and extended recall time

Subjects learned a single A-B or successive A-B, A-C lists and were then tested for recall of the first-list materials after 1 week. In the MMFR conditions a typical MMFR task was administered. In the extended conditions Ss were allowed 1 week to try to recall the materials. Both recall techniques produced significant RI. Ss in the extended conditions were able to recall a significant number of additional items during the week provided for recall. In agreement with the contention that competition has little effect upon recall, and in contrast to the assumption that extended recall might allow for the dissipation of competition effects, equal amounts of RI were obtained with the two recall techniques.

Retroactive effects in S1-R2 and S2-R1 paradigms

A preliminary study was conducted to indicate whether two transfer surfaces described by Houston (1964; 1965) would be useful in predicting retroactive effects. First-list retention was measured in five two-list paradigms (A-B, B-A; A-B, C-A; A-B, C-D; A-B, B-C; A-B, B-Ar) and a single-list control condition. With one major exception the pattern of the results conformed to the predictions of the surface. The exception was the A-B, C-A paradigm which produced much poorer first-list recall than that predicted by the surfaces. While not overwhelmingly convincing, these data suggest that the surfaces may contain some accurate predictions concerning retroactive effects.

A repaired S1

A surface was described which summarized the transfer effects expected in two-list paradigms varying in S1-R2 and S2-R1 similarity when the first-list units were repaired in the second list. According to the surface, negative transfer should increase as S2-R1 similarity increases. This increase should be greater when S1-R2 similarity is high than when it is low.

A repaired S 1 -R 2 and S 2 -R 1 transfer surface

A surface was described which summarized the transfer effects expected in two-list paradigms varying in S1-R2 and S2-R1 similarity when the first-list units were repaired in the second list. According to the surface, negative transfer should increase as S2-R1 similarity increases. This increase should be greater when S1-R2 similarity is high than when it is low.