Comparison of a Rasch model scale and the grade-equivalent scale for vertical equating of test scores

Abstract

Wright (1977) outlined procedures for equating tests and test scores using the Rasch model. This paper summarizes the results of a study in which those Rasch model procedures are used to calibrate and then to link vertically six levels of the Reading Comprehension Test of the Iowa Tests of Basic Skills. The derived Rasch ability scale estimates are then compared to norm-referenced, grade-equivalent scale estimates for scores across the test levels. The results of these comparisons suggest that where discrepancies in the two scales emerge, the more accurate and perhaps more useful measure is provided by the Rasch scale.