Norman J. Slamecka’s research while affiliated with University of Toronto and other places

Three experiments investigated the effects of a prior test trial on the subsequent long-term forgetting of paired-associates lists. The pattern of findings was consistent with a learning hypothesis that views a prior test as promoting the overlearning of remembered items and as mediating an efficient learning strategy on the next study trial, but leaving the forgetting rate intact. The learning hypothesis was upheld when pitted against the contrasting predictions of a retrieval-based notion, and it also provided an interpretation of extant data from the method of successive relearning. The main conclusion was that, although a prior test has substantial performance consequences, there is no evidence that it influences the rate of forgetting. (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Four experiments on the generation effect of free recall were reported. Experiment 1 used a within-lists design that showed the generation effect with unilingual, but not bilingual, pairs. Experiment 2 had a between-lists design and found no effect with either kind of pairs. Experiment 3 contracted the two design types with unilingual materials, and obtained generation effects only for the mixed-list arrangement. Experiment 4 prevented displaced rehearsal in a mixed-list arrangement, and found no generation effects. It was concluded that the generation effect of recall is an artifact of selective displaced rehearsal that strengthens generated items at the expense of read items. It was also argued that these data render extant theories of the effect untenable.

Discusses G. R. Loftus's (see record 1986-05498-001) proposal for comparing relative forgetting rates by evaluating the horizontal interaction rather than the standard vertical interaction, as used by the present author and B. McElree (see record 1984-05823-001) in their study of forgetting of verbal lists as a function of their degree of learning. It is maintained that the proposed method entails an unavoidable confounding with the ages of the lists being compared, such that one list is measured at an earlier point in its forgetting curve than the other. This prevents an unequivocal assessment of the effect of the treatment variable of interest. The vertical interaction appears to be free of artifact. It is concluded that although the horizontal interaction has the advantage of circumventing the scaling problem, this is more than outweighed by the disadvantage of the list-age confound. (7 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Responds to comments by W. K. Estes, W. Kintsch, G. Mandler, B. B. Murdock, T. O. Nelson, E. Tulving, and R. K. Young (see PA, Vol 73:10996, 11006, 11011, 11012, 11014, 11026, and 11032, respectively) on the contributions of H. Ebbinghaus to the history of psychology. Some of the ideas expressed by these authors include the following: (1) Because he was a true pioneer, it was argued that Ebbinghaus's influence had to be, on balance, positive. (2) It was maintained that there is no sharp line of demarcation between association and structure but rather that the presence of an association always implies structure. (3) It was reaffirmed that Ebbinghaus's passing glance at the manifestation of recollective experience suggests that this subjective indicant of memory was epiphenomenal with respect to an objective indicant. (4) It was emphasized that Ebbinghaus based his seminal work on the measurement of implicit memory, which is now undergoing a resurgence of experimental concern with respect to notions of dissociation and separate memory systems. (14 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

In recognition of the 100th anniversary of the publication of H. Ebbinghaus's monograph Memory: A Contribution to Experimental Psychology (1885 [1964]), the present author comments on various aspects of the monograph in light of subsequent experimental and theoretical developments. Specific topics discussed are remote associations, the chaining hypothesis, the length–difficulty relation, the long list–short list problem, successive relearning, the law of contiguity, recollective experience, ecological validity, and sources of bias. In a general evaluation of the Ebbinghaus legacy, it is concluded that in spite of extensive changes that have occurred in the field, Ebbinghaus's work still stands as a model of excellence. (66 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Three experiments, with 276 undergraduates, examined how degree of learning affects normal forgetting. Exp I varied learning of categorized lists and tested retention at 3 intervals (immediately and 1 or 5 days after presentation). Across all measures, study trials affected intercepts but not slopes of forgetting functions. Exp II varied learning of paired-associate lists and tested retention at the same 3 intervals. Across all measures, trials influenced intercepts but not forgetting slopes. Exp III varied learning of sentence lists and tested verbatim and gist memory at the same intervals. Again, trials affected intercepts but not slopes. Results suggest that the forgetting of verbal lists is independent of their degree of learning. No current theories of memory predict these outcomes, but neither does the pattern of results disconfirm any theory. It is argued that present memory theorizing neglects almost entirely the central problem of normal forgetting. (36 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

The memorial consequences of generating and reading nonwords were compared under a variety of input and test conditions. The findings obtained refuted the notion that the generation effect was due to inherent differences in the generate and read processes per se. The results suggest that generating yields superior retention only when semantic memory is engaged at the time of study. Explanations for the generation effect consistent with this conclusion were discussed.

Reports on 5 experiments with 96 undergraduates, comparing memory for words that were generated by the Ss themselves with the same words when they were simply presented to be read. In all cases, performance in the Generate condition was superior to that in the Read condition. This held for measures of cued and uncued recognition, free and cued recall, and confidence ratings. The phenomenon persisted across variations in encoding rules, timed or self-paced presentation, presence or absence of test information, and between- or within-Ss designs. The effect was specific to the response items under recognition testing but not under cued recall. A number of potential explanatory principles are considered and their difficulties enumerated. It is concluded that the generation effect is real and that it poses an interesting interpretative problem. (23 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Performed 3 experiments with 80 university students to test the prediction that the serial to derived paired-associates transfer task cannot be expected to yield more than relatively moderate degrees of transfer, primarily because one aspect of serial performance (i.e., utilization of response-produced cues) cannot transfer productivity to the 2nd task. Exp I tested serial to derived paired-associates transfer while minimizing biases from response guessing and backward associative interference; moderate transfer was found. Exp II modified the paradigm to encourage the productive transfer of response-produced cues; substantial positive transfer resulted. Exp III was designed to detect the utilization of response-generated stimuli in the absence of confounding with position information. It is concluded that serial anticipation learning is guided by a stimulus complex which includes the prior item on the memory drum together with cues produced by the correct prior anticipatory response. (23 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Double-function list acquisition was analyzed in three experiments which used mixed-list designs and forced-choice recognition tests. Findings showed that specific associative interference occurred both during learning and performance phases, that its source was limited to the immediate backward associate during the learning phase, but that it also involved items further removed in both the forward and backward directions during the performance phase, probably by way of mediated chaining.