Files in this item
|(no description provided)|
|Title:||A Conceptual Framework for Investigating Test Item Performance With the Mantel-Haenszel Procedure|
|Author(s):||Ryan, Katherine Elizabeth|
|Doctoral Committee Chair(s):||Linn, Robert L.,|
|Department / Program:||Education|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Subject(s):||Education, Tests and Measurements|
|Abstract:||Standardized tests are frequently used to assess the educational achievement of students in specific content areas (e.g., reading comprehension or mathematics). However, standardized tests have been highly criticized as unfair with little scientific objectivity when assessing the performance of particular racial or minority subgroups. Proponents counter that elaborate procedures involving test specifications, item reviews and sophisticated statistical analyses have been developed to ensure the most valid, unbiased test possible. However, differential subgroup performance continues. Whether these differences can be attributed to educational instruction, educational opportunity, and/or some intrinsic properties of achievement test items has not been resolved.
This dissertation examined the relationship between differential item performance by eighth grade Hispanic and black students and specific item characteristics on the Second International Study of Mathematics. In addition, it investigated whether the identification of differentially performing items can be improved by controlling for the multidimensionality of a matching criterion by controlling on an additional criterion. The performance of the Mantel-Haenszel procedure for various sample sizes of different test takers and different sets of items was examined.
Test items were classified within a six category system drawing from Cummins (1983). The findings suggested that further refinement of the classification is necessary. In addition, the results indicated larger samples of test takers are needed to obtain stable estimates of the Mantel-Haenszel indices as well as to determine whether differential item performance can be improved by controlling for the multidimensionality of a test.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1988.
|Date Available in IDEALS:||2014-12-15|