A study of selected respone aberrance indices as alternatives to the standard error of measurement in interpreting individual text scores: y Judy A. Shishido

Publication Year:
Usage 29
Abstract Views 17
Downloads 12
Repository URL:
Shishido, Judy A.
Educational tests and measurements -- Evaluation
thesis / dissertation description
The purpose of this study was to discover any relationships between three aberrant test response indices, the Modified Caution Index (MCI), the Person Average R (PAR),and the item-response theory-based standardized logistic maximum likelihood function (Z3); and four hypothesized reasons for aberrant response patterns, curricular differences, differences in test-taking skills, differences in motivation, and differences in consistency of academic performance. Two secondary purposes of this study were to evaluate the effectiveness of the aberrance indices as aids in interpreting test scores, and to investigate the occurrence of sandbagging on a placement test. Data from the University of Hawaii at Manoa's Japanese Language Placement Test was used, along with survey data obtained from students taking the Placement Test, students enrolled in UH Manoa Japanese language courses, Hawaii high school Japanese language teachers, and UH Manoa Japanese language instructors. Relationships were found between MCI and PAR and curricular differences and differences in consistency of academic performance, and between all three indices and differences in motivation. It was estimated that between four and ten percent of test takers engaged in sand bagging on the Placement Test The use of either MCI or PAR was recommended as aids in interpreting Placement Test scores.