Evaluating forgetting curves psychologically.

Abstract

A theoretical treatment of forgetting is presented which uses scaled variables, provides a new definition of independence of rate of forgetting from initial level of learning, and provides an appropriate statistical test of that independence based on functional measurement methodology. It is shown that Loftus's (1985a) horizontal interaction method leads to conclusions that are at odds with existing theory, and his conclusion that forgetting is slower for higher degrees of original learning is weakened thereby although it is consistent with his method. Arguments are given for preferring the method presented here. Slamecka and McElree's (1983) approach is rejected, but their conclusion that forgetting rate does not depend on original learning is supported.I am indebted to Donald Bambcr, Geoffrey Loftus, Sandy Pollatsek, Keith Rayner, Norman Slamecka, and two anonymous reviewers for helpful comments on earlier drafts.

Perceptual and conceptual priming in a semantic reprocessing task

“…Although the difference between the consistent and inconsistent repetitions appears fairly constant in Figure As noted by several researchers, conclusions drawn from the comparison of forgetting rates when there are different initial levels oflearning (in this case, priming) requires an assumption that the dependent measure is linearly related to the relevant psychological dimensionin this case, changes in memory strength (e.g., Bogartz, 1990;Loftus, 1985;Slamecka, 1985;Wixted, 1990;Wixted & Ebbesen, 1991). Such an assumption is difficult to justify, and there appears to be no easy way to resolve this scaling dilemma.…”

One hundred years of forgetting: A quantitative description of retention.

“…The particular question that has generated the most debate in recent years is whether retention functions differ with the amount of learning (Bogartz, 1990a(Bogartz, , 1990bLoftus, 1985aLoftus, , 1985bLoftus & Bamber, 1990;Slamecka, 1985;Slamecka & McElree, 1983;Wixted, 1990). Slamecka and McElree (1983) examined the literature on the effects of degree of learning on retention and found that A fair summary of the appearance of forgetting curves from all sources reviewed is thai they either had equal slopes or slopes that varied directly with intercept levels.…”

“…Different choices of a function could lead to different conclusions for the same data. Nonetheless, in most of the debate so far, regardless of theoretical positions on how retention should be measured, the equation has either been assumed to be y = b'g~"" or has been chosen to fit the particular data set (Bogartz, 1990a(Bogartz, , 1990bLoftus, 1985aLoftus, , 1985bLoftus & Bamber, 1990). If a retention function (or a set of retention functions) could be chosen that fit a class of retention data, one could return to this question using that function.…”

“…Bahrick (1965) has noted advantages for plotting functions in units of variance, Slamecka and McElree (1983) have argued for a standard analysis of variance (ANOVA) technique, Loftus (1985aLoftus ( , 1985bLoftus & Bamber, 1990) has developed a theoretical framework based on comparing functions horizontally instead of vertically, and Bogartz (1990aBogartz ( , 1990b has considered splitting the observed function into underlying unobserved components corresponding to assumed memory processes. The simplest approach, and the one adopted here, is to fit the empirical data using standard curve-fitting techniques first and to consider complications only if necessary.…”

Forgetting functions

“…A different conclusion follows from rate of forgetting estimated directly as a parameter of a fitted function. Bogartz (1990), Loftus (1985), and Loftus and Bamber (1990) have debated the issue of how to assess changes in forgetting rate independently of the level of original learning. Wixted (1990) concluded that the best approach was to fit a mathematical function to the data relating performance to retention-interval duration and to use the slope parameter as a measure of rate of forgetting.…”