The negative generation effect: Delineation of a phenomenon

Abstract

In four experiments employing between-list designs, generation was found to have negative effects on free recall of word pairs and on cued recall of the second word. In addition, generation had negative effects on measures of word-pair integration and on clustering in recall. In contrast, positive effects of generation were found on free recall of second words alone, and on a recognition test for memory of the second word. It was concluded that in between-list designs, generation led to greater individual-item processing of the generated term than reading, but this processing occurred at the expense of processing the relation between the words in a pair and processing the relations between different pairs in a list. Slamecka and Graf (1978) demonstrated that recognition of the second word from word pairs was greater if subjects were required to fill in missing letters from that word than if subjects were simply asked to read the word pairs. This phenomenon has been called the generation effect. The generation effect has been demonstrated with word pairs

The generation effect and the modeling of associations in memory

“…It was predicted that under these conditions, generation facilitation should be reduced, particularly for free recall (Schmidt & Cherry, 1989;Slamecka & Kasaiti, 1987). Precisely this pattern of results was shown in Experiment 2.…”

“…Previous experiments have shown that generation effects for free recall are sometimes absent or even negative (Hirshman & Bjork, 1988;Schmidt & Cherry, 1989;Slamecka & Kasaiti, 1987). Many, though not all, ofthese failures to show generation facilitation have manipulated encoding conditions between subjects or have used intentional learning procedures, or both.…”

A test of resource-allocation explanations of the generation effect

“…A more complete explanation of the generation effect must rely on interactions between list structure, types of information encoded, and the nature of the memory test (McDaniel et aI., 1988). In a between-subjects design, generation appears to increase individual-item processing at the expense of whole-list organization (Schmidt & Cherry, 1989). On memory tests in which list organization is important (e.g., free recall), the increased individual-item processing and decreased relational processing may lead to no net effect of generation (Slamecka and Katsaiti, 1987), or to a negative generation effect (Schmidt & Cherry, 1989).…”

“…In a between-subjects design, generation appears to increase individual-item processing at the expense of whole-list organization (Schmidt & Cherry, 1989). On memory tests in which list organization is important (e.g., free recall), the increased individual-item processing and decreased relational processing may lead to no net effect of generation (Slamecka and Katsaiti, 1987), or to a negative generation effect (Schmidt & Cherry, 1989). In contrast, on recognition (Schmidt & Cherry, 1989;Slamecka and Graf, 1978) or cued recall tests (Hirshman & Bjork, 1988;McDaniel et aI., 1988), list organization is less important, and positive effects of generation are found even in betweenlist designs.…”

Can we have a distinctive theory of memory?

“…Similar frameworks have been adopted to explain bizarre imagery (McDaniel & Einstein, 1986), the concreteness effect (Marschark & Hunt, 1989), and the generation effect (Schmidt & Cherry, 1989). Furthermore, the hypothesis does not seem at odds with Eysenck's (1979) ideas.…”