The effects of inter-trial spacing in paired-associate learning

Abstract

First few paragraphs: Paired-associate recall following two presentations of a pair to be remembered is heavily influenced by the spacing (in terms of intruding trials involving different pairs) between the two successive presentations. In particular, subsequent recall performance improves as the spacing between two successive presentations of a pair increases, at least up to some optimal interpresentation spacing interval. This effect is known as the spaced practice improvement (or SPl) effect, and is clearly of fundamental importance to our understanding of the relationship between repetition and learning. However, most of the recent research on the SPI effect has involved free-recall and Brown-Peterson paradigms, and there are grounds for suspecting that the SPI effect obtained with paired associates may have a different underlying rationale to the SPI effect observed in these other paradigms. Although the extant data strongly suggest that pairs held in short-term memory al the time of their second presentation receive little or no benefit from that re-presentation, there has been no systematic work attempting to relate the effectiveness of a re-presentation with both interpresentation spacing and with the state of learning oi a pair at the time of the re-presentation. This thesis was designed to investigate t is relationship in an effort to e ivs constraints on an adequate theory of the SPI effect in paired-associate learning beyond those imposed by prior research. To this end, three experiments were conducted, each of which employed a variation of the study-test, continuous paired- associate (CPA) paradigm. The basic condition common to all three experiments may be depicted as [equations displayed] where P1 and P2 are presentations of a pair to be remembered, T1 and T2 are tests of the pair, i represents the spacing between the two successive presentations in terms of intruding trials involving other pairs, and there were always 8 such trials between the second presentation P2 and the final test of a pair, T2. It will be noticed that T1 always immediately preceded P2, so that T1 performance would give some insight as to the state of learning of a pair on entry to P2.