テストが再学習の成果に及ぼす影響

抄録

This study is a part of a serial, studies on the effects of testing upon learning achievement, and the experimental designes are based on a hypotheses as follows:
Testing on the materials which the pupils previously learned should make them recognize what they have mastered, and what that they have not mastered of the materials. Then, the following assumptions can be taken for granted:
1) A failed part of the learned materials as recognized by a pupil would motivate him to relearn the part.
2) Relearning the materials, they would more effectively learn it with efforts or attention.
3) The memory traces which remained from original learning will be reinforced by relearning and retesting.
The main purpose of this experiment is to confirm the assumptions of 1) and 2).
Using “Two-number figures” and “nonsense syllable” as learning materials, and using “the control group method” and “the rotation method,” four experiments in all were gived. The details of the experimental procedures differed on each experiment. Fig. 1, 2, 3, and 4 show the processes of four experiments. And the essential process in common to all those experiments was: 1) to let two equivalentgroups equivalentgroups: the experimental and the controlled, learn identical materials, 2) to give a test on the materials to the experimental groupcertain days after they were presented, 3) to let both the groups relearn the original materials later and then 4) to examine the influences of the test given upon the relearning products, comparing two groups.
The effect in answering to the test given was examined by means of comparing the differenecs in the achievement of two kinds of task: a recall of the materials right after relearning, and that of certain days after the first recall. Achievement of each experiment of 1, 2, 3, and 4, are shown in each table of 1, 2, 3, 4, 5, and 6 respectively. Table 7 is the general picture of all experiments.
In Table 7, there are presented 12 indicies in all in order to closely examine the validity of the hypotheses mentioned above. Six of the indicies (*marked in Table 7) positively support the validity of our hypotheses and four of them have an inclination toward agreement with our hypotheses although they are not so on an significant level. The other two indicies (underlined in Table 7) are against to our hypotheses, but the difference is small enough to be ignored.
Thus, all the results of this experiment prove that our working hypothese are valid.