処理のレベルと再認記憶における文脈効果との関係

猪木 省三

doi:10.4992/jjpsy.50.241

LEVELS OF PROCESSING AND CONTEXT EFFECTS IN RECOGNITION MEMORY

SHOSO INOKI

Author information

JOURNAL FREE ACCESS

1979 Volume 50 Issue 5 Pages 241-248

DOI https://doi.org/10.4992/jjpsy.50.241

Details

Published: January 20, 1980 Received: January 30, 1979 Available on J-STAGE: July 16, 2010 Accepted: - Advance online publication: - Revised: -
Correction information
Date of correction: July 16, 2010 Reason for correction: - Correction: ABSTRACT Details: Wrong : The present study was designed to examine whether levels of processing at input have an effect on context effects in recognition memory when the context is defined as the word paired with the to-be-remembered (TBR) word.
In Exp. I levels of processing were manipulated by presentation rates. A 2×2 factorial design was used. The first factor involved two different rates of presentation: 2sec and 4sec per pair. The second factor dealt with whether contexts were unchanged or changed from study to test. The first was a between-subjects factor, and the second was a within-subjects one.
The subjects were 32 university students. They were presented with a study list of 50 word-pairs, and each of them consisted of a TBR word and a context word. The first 5 and the last 5 pairs in the list were eliminated from the recognition test to reduce primacy and recency effects. Following the presentation the recognition test was given. The test consisted of 40 TBR and 40 distractor words, each paired with a context word. Half of the 40 TBR words were paired with changed contexts, while the other half with unchanged ones.
The results of Exp. I are shown in the left half of Table 1. ANOVA of these results revealed no differential effects of presentation rates on context effects. Another analysis, using the relative performance (the ratio of the performance in the changed context to that in the unchanged context), showed no difference in context effects between presentation rates. The results of a supplementary experiment, in which the presentation rates were 3sec and 6sec per pair, are shown in the right half of Table 1. These results show a pattern similar to that in Exp. I.
In Exp. II levels of processing were manipulated by orienting tasks instead of presentation rates. For the semantic task subjects were required to rate the pleasantness of word-pairs, while the non-semantic task was to count the number of particular letters in word-pairs. Other details of the method employed were the same as in Exp. I.
The results of Exp. II are shown in Table 2. ANOVA of these results showed a context effect×orienting task interaction at the significance level of approximately .05. The analysis of the relative performance, however, showed no difference in context effects between semantic and non-semantic orienting tasks.
The findings described above indicate that levels of processing had no effect on context effects in recognition memory, when words were used as the contexts of TBR words. In the previous experiments where sentences were used as contexts the present author found that context effects were affected by levels of processing, and the result was regarded as supporting the explanation of context effects in terms of the change of subjects' semantic interpretations of TBR words. The findings of the present study, however, cannot be explained in the same manner, because even the non-semantic processing led to context effects. In the unchanged context both the TBR and the context words in the test list were the same as in the study list, and could be used as retrieval cues, whereas in the changed context only the TBR word served as cue. Therefore, it was suggested that context effects in the present experiments were caused by the difference in number of available cues between the unchanged and changed context conditions.
Right : The present study was designed to examine whether levels of processing at input have an effect on context effects in recognition memory when the context is defined as the word paired with the to-be-remembered word. Levels of processing were manipulated by presentation rates (2sec vs. 4sec per pair) in Exp. I, and by orienting tasks (semantic vs. non-semantic) in Exp. II. Context effects were found irrespective of levels of processing in both experiments. It was suggested that context effects in the present experiments were caused by the difference in number of available retrieval cues between the unchanged and changed context conditions, not by the change of subjects' semantic interpretations of to-be-remembered words.

Abstract

The present study was designed to examine whether levels of processing at input have an effect on context effects in recognition memory when the context is defined as the word paired with the to-be-remembered word. Levels of processing were manipulated by presentation rates (2sec vs. 4sec per pair) in Exp. I, and by orienting tasks (semantic vs. non-semantic) in Exp. II. Context effects were found irrespective of levels of processing in both experiments. It was suggested that context effects in the present experiments were caused by the difference in number of available retrieval cues between the unchanged and changed context conditions, not by the change of subjects' semantic interpretations of to-be-remembered words.

Corresponding author

Correction information

Register with J-STAGE for free!