Files in this item
|(no description provided)|
|Title:||The Effect of Explicitly and Implicitly Presented Rhetorical Functions on The Comprehension of Scientific Discourse: A Multitrait-Multimethod Construct Validation Study of Est Reading Comprehensions|
|Department / Program:||Education|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Subject(s):||Education, Tests and Measurements|
|Abstract:||The purpose of this study was to examine whether the ability to comprehend an explicitly presented rhetorical function and the ability to comprehend an implicitly presented rhetorical function in scientific discourse constituted two distinct traits. This study also examined the relative effect of these two traits and several testing methods on performance of students on reading comprehension tests.
A multitrait-multimethod design with two traits and four testing methods was used. Testing methods employed comprised two test formats, cloze test and multiple-choice test, and two rhetorical functions, argumentation and classification. All tests were given to 114 non-native students of English at the University of Illinois at Urbana-Champaign. In addition to examining the patterns of correlations among measures in the multitrait-multimethod matrix, this study used confirmatory factor analysis to examine the factor structure underlying the observed measures and to test several hypothesized primary factor models.
A comparison of the relative magnitudes of the correlations in the multitrait-multimethod matrix indicated a strong testing method effect on all of the measures, as well as low discriminant validity. The confirmatory factors analysis indicated that the model of best fit included two distinct but highly correlated traits. In interpreting the relative importance of the trait, test format, and rhetorical function factor loadings in this model, information from a questionnaire on subjects' familiarity with the content of the passages was also considered. There was a significant correlation between the levels of familiarity and the performance on the tests. Factor loadings indicated that the less familiar passage provided better measures of the traits than did the familiar passages. In addition, there were strong interactions among the familiarity with the content, the test formats, and the rhetorical functions of a passage. The results of this study suggest that construct validation can provide important information for test users. This information may not only prevent the careless misuse of tests, but also provide a better understanding of the nature of the traits underlying the tests and the characteristics of testing methods used.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1983.
|Date Available in IDEALS:||2014-12-15|