読解力を測定するテストの一つにリコール（recall）テストがある。読解におけるリコールには採点規準を厳密にすべきかどうかという評価の問題がある。もとの文章の内容を忠実に復元させる厳密な採点基準と大体の情報が再生されていれば正解とするゆるやかな採点基準がある。本研究では，高校生（特に上位校に在籍する高校生）を対象に、リコール・プロトコル（recall protocol）の異なった採点基準による採点（厳密な採点基準とややゆるやかな採点基準）が，（1）テクスト全体，（2）テクスト中の重要な情報（主題文全体），（3）各パラグラフ全体，（4）各パラグラフの主題，の再生量にどのような影響を及ばすかを検証した。その結果，（1）2種類の採点基準の相関は極めて高かった。すなわちテクスト全体のリコールでは，2種類の採点基準の相関係数は．96で，重要な情報量（主題文全体）のリコールでは，81であった。また（2）各パラグラフ全体及び各パラグラフの主題の再生率の順位（高低）のパターンが，採点基準の違いによって，異なることが明らかになった。

This study investigates the effects of two different scoring criteria on the written recall of an expository text. Using both a strict and a loose criterion of recall, the written recalls of 53 Japanese EFL high school students (focusing on students with high proficiency) were scored. The strict criterion disallowed distortions and additions, though it allowed paraphrases. The loose criterion allowed some distortions for the original meaning, i .e. , cases in which most of the meaning was correctly recalled. The results revealed that the correlation between the strict and the loose criterion was extremely high for the total number of idea units recalled correctly ( r = .96, p<.01). This finding for Japanese high school students supported the very high correlation found with Japanese graduate students in Hirano (1996). For the total number of main ideas recalled correctly, the correlation was also high (r = .81, p<.01) . In addition, the criterion of scoring recall protocols applied proved to be a significant factor which affected the comparisons of the recall of each paragraph and of each paragraph main idea. This finding reaffirmed the results of Hirano (1996). That is, significant differences were found between the strict and the loose criterion in terms of the percentage of correct recall across paragraphs and paragraph main ideas. For paragraph recall, there were significant differences between all four adjacent paragraphs recalled in the strict criterion: Paragraph 1 > Paragraph 2 > Paragraph 3 > Paragraph 4. In the loose criterion, however, there were no significant differences for all paragraphs: Paragraph 1 > Paragraph 2 = Paragraph 3 > Paragraph 4. For the recall of each paragraph main idea, the pattern that emerged was different for each criterion: for the strict criterion, Main Idea 1 = Main Idea 2 > Main Idea 3 > Main Idea 4; for the loose criterion, Main Idea 1 > Main Idea 3 = Main Idea 4, Main Idea 2 > Main Idea 4. (Main Idea 1 = Main Idea 2, Main Idea 2 = Main Idea 3)