Evaluating Student Multiple-Choice Responses

Publisher Website
Google Scholar
Abstract
The effects of coded (type K) and free (type X) formats on student responses to identical multiple-choice items were com pared. Two subtests (A and B) each of 20 multiple-response items were emplo ved The items of the two subtests were placed in both K and X response formats pro ducing two examinations (i.e., exam I subtest A in type K format and subtest B in type X format; exam II. subtest A in type X format and subtest B in type K format) which were randomly distributed to 155 premedical students enrolled in an advanced human physiology course. Analyses of student performance led to four observanons. (1) the type K items significantly elevated student achievement scores due to a cueing effect inherent in the format, (2) the "poorer" students' scores were elevated more than the "better" students' scores, as evidence by a signifi cant negative relationship between the degree of cueing and examinee perfor mance; (3) the item format and scoring system (dichotomous or semicontinuous scales) affected the mean and variance of examinee scores; (4) the reliability (alpha coefficient) of items in the type X format tended to be statistically higher than that of items in the type K format. It was con cluded that an examination constructed in the type X format is a superior evalua tion instrument when compared to its parallel type K format

This publication has 3 references indexed in Scilit: