Examining the Difficulty of Large-Scale Mathematics Assessment Items via Explanatory Item Response Theory Models

• In Event: New Developments and Applications in Cognitive Diagnosis Models

Abstract

Examining the diverse sources of item variation allows for an increased understanding of an assessment item’s difficulty. Information obtained from analyses of this type aid in item and test construction, as well as inform classroom instruction. This study examined the relationship between process skills and item difficulty across different versions of a large-scale mathematics assessment and between different student groups. Using explanatory item response theory models, item difficulty was modeled using the process skills associated with the items. Preliminary results suggest that process skills do influence item difficulty. However, their impact on difficulty (i.e., increasing or decreasing it) varies across grades.

Authors

• Kristin M. Morrison, Georgia Institute of Technology

• Robert Andrew Schwartz, Pearson