The Effects of Lexical Meanings on Cloze Scores at the Sentential and Textual Levels
THE RESEARCH PRESENTED HERE explores the effects of a knowledge of lexical meanings on cloze scores at the sentence and text levels. Four groups of sujects were used in testing. The first two groups, one of students majoring in English and another of non-English majors, took scrambled-type and sequential-type cloze tests with the meanings of vocabulary items provided (hereafter, respectively EL group and NE group). A further two groups, one of English majors and the other of non-English Majors, were not presented with lexical meanings for the same tests (hereafter, respectively E group and N group). Three tests were given to all groups: a vocabulary test, a scrambled-type cloze test, and a sequential-type cloze test.
A 2 x 2 x 2 ANOVA (English majors vs. non-English majors, lexical meanings presented vs. not presented, and sentential vs. textual) was used to analyze the test scores. Correlations between sentence scores (scrambled-type cloze scores), cohesion scores (scores obtained by subtracting sentence scores from text scores) and text scores (sequential-type cloze scores) were computed for each group (see Tables 3 and 4). Results showed that all the main effects were significant, but, apart from the triple interaction, the interactions were not significant. Sentence scores demonstrated a significant correlation with text scores for the NL, N, and EF groups. In the case of the E group, however, text scores correlated not with sentence scores but with cohesion scores, the results of which were interpreted by means of a kind of floor effect in the scrambled-type cloze test.
This type of study indicates, first, that a familiarity with the meanings of vocabulary items used sentential and textual clozing produces a relatively small though statistically significant effect. Moreover, the present paper points out that the difference in scores between the students majoring in English and those concentrating on other subjects proved smaller than expected; consequently comparison of these two categories as such is not revealing. Finally, this research suggests the importance of the phrase in reading comprehension. When the responses of all four groups of subjects were tallied to highlight the relative difficulty of various cloze items, the results showed that the difficulty of a vocabulary unit hardly depends on whether or not it functions across sentences as a cohesive item. The comprehensibility of an item appears independent of sentence complexity as represented by the simple vs. complex sentence. Instead, the results of this study support the conclusion that the difficulty of a cloze item for the reader is related to the "difficulty" of the phrase in which it stands. This conclusion underlines the value of a phrase-oriented approach both in language learning and in language research.
|date of issued||
Departmental Bulletin Paper
Departmental Bulletin Papers
Graduate School of Integrated Arts and Sciences