Using derived schemas and scripts: impact on recall, metacognition, and "consumer satisfaction"

Abstract

The present study explores the use of consciously derived schemas and scripts to enhance recall of a scientific passage. Five experimental groups were used: Groups 1, 2, & 3 all received expert training to use an expert-derived schema and script, but differed in passage supplement: Group 1--completed expert supplement, Group 2--blank expert supplement, and Group 3--no supplement, but was told to use the training to assist in notetaking. Group 4 derived their own schemas and scripts and used them as notetaking tools, and Group 5 was told to study the passage in their usual manner, and received expert training between the first and second recall tests. The groups studied the target passage, and recalled the information on free and cued recall tests two days later. Vocabulary level, prior knowledge, field independence/dependence, and analogical reasoning were also assessed. Four statistical analyses were conducted. First, a MANOVA on the free and cued recall tests revealed no significant differences between the five experimental groups. Second, regression analyses indicated that vocabulary level and prior knowledge predicted performance in the schema-explicit conditions (1 and 2), while only prior knowledge predicted performance in the schema-implicit conditions (3 and 4), and only vocabulary level predicted performance in the no schema at encoding condition (5). Third, Fisher's z comparisons of correlations between performance predictions and performance indicated no difference in predictions on the free recall test. However, on the cued recall test, the student-generated schema group (4) was better able to predict their performance than the other groups. Last, analysis of a post-experimental questionnaire which measured subjects' opinions of the helpfulness of the training ("consumer satisfaction") was conducted with prior knowledge as the covariate and experimental group as the independent variable. This test indicated that the student-generated schema group (4) was more positive about the experiment than were the other groups. Cognitive overload and heterogenous group composition are offered as explanations for the surprising and disappointing results. But, the student-generated schema group deserves further consideration: their performance was consistently higher on the recall tasks, they had better metacognitive knowledge (performance prediction), and they liked the training better than the other groups.