Levels of processing and the effect of spacing upon judgements of frequency
AbstractThe term spacing effect refers to the empirical fact that items which are repeated with few other items intervening between the repetitions are remembered worse than items which are repeated with relatively more interventions between the repetitions. The purpose of the set of spacing experiments reported here is to infer a cause of the spacing effect in a judgment of frequency paradigm by trying to discover conditions which will remove it. -- Following Experiment 1, which compared continuous judgments of frequency of words made on the learning trial with terminal judgments of frequency made at the conclusion of the learning trial, it was inferred that the spacing effect arose from one or more of three possible origins: (1) a true memory deficit for massed repetitions together with a biassed tendency to overestimate the frequency of words repeated at non-zero spacings; (2) the use of different strategies on continuous and terminal judgments; (3) the deficient processing of repetitions at low values of spacing relative to repetitions at higher values. -- Experiments 2 and 3 found evidence which was irreconcilable with the first two of these three possible origins of the spacing effect. Experiments 4 and 5 were then carried out to test the relevance of the third possibility. Specifically, the levels of processing hypothesis was contrasted with the variable contextual encoding hypothesis. The former claims that the spacing effect arises because repetitions at long spacings receive deep, reconstructive processing while those at short spacings receive shallow scanning processing. The contextual encoding hypothesis, which enjoys some support elsewhere, attributes the spacing effect to the greater variability among the contents of the repetitions which occur at long spacings. The evidence from Experiments 4 and 5 generally supported the levels of processing hypothesis but in addition indicated that the encoding context is a factor which interacts with the measure of retention. Also the nature of the posited scanning and reconstructive processes still remains a mystery.
Rose, Robert James <http://research.library.mun.ca/view/creator_az/Rose=3ARobert_James=3A=3A.html> (1978) Levels of processing and the effect of spacing upon judgements of frequency. Doctoral (PhD) thesis, Memorial University of Newfoundland.