Detecting Language Differential Item Functioning, Item Bias, and Item Impact: A Study Using Propensity Scores

• In Event: 34.079 - Roundtable Session 14
  In Roundtable Session: 34.079-11 - Ordinal and Binary Data: From Differential Item Functioning to Classification and Regression Tree

Abstract

Through our present research, a unified methodology is applied for detecting differential item functioning, item impact, and item bias in items. The methodology utilizes the technique of matching on balance scores, specifically propensity scores, to help rule out alternative explanations for the causal inferences embedded in the definitions of item impact and item bias. The methodology was applied to data from TIMSS 2011 Canadian student’s eighth grade science examination. The grouping variable was language with English test takers as the reference group and French as the focus group. Results showed that, of the 28 items studied, 12 items were classified as having no impact, 3 items were identified as having impact, and 13 items were classified as biased.

Authors

• Danjie Zou, The University of British Columbia

• Amery D. Wu, University of British Columbia

• Yan Liu, The University of British Columbia

• Bruno D. Zumbo, The University of British Columbia