skip to main content skip to footer

Identifying Differentially Functioning Items in the NTE Core Battery DIF NTE

Author(s):
McPeek, W. Miles; Wild, Cheryl L.
Publication Year:
1992
Report Number:
RR-92-62
Source:
ETS Research Report
Document Type:
Report
Page Count:
113
Subject/Key Words:
Differential Item Functioning (DIF), Item Analysis, Mantel-Haenszel Technique, NTE Core Battery, National Teacher Examinations (NTE), Racial Differences, Sex Differences

Abstract

The purpose of this study is to evaluate the use of the Mantel- Haenszel statistic as a methodology for identifying differentially functioning items on the NTE Programs Core Battery tests. Retrospective analyses of data collected over a three- year period are reported for Black/White, Hispanic/White, and female/male comparisons. The three tests that comprise the Core Battery are included in the analyses--Communication Skills, General Knowledge, and Professional Knowledge. Data concerning the replicability of the results, the number of items identified as differentially difficult, and the characteristics of the test questions associated with differential difficulty are discussed. Replicability of DIF values from two administrations based on correlational data was modest to high, with greatest replicability found in the largest samples. Although the replicability is modest to high when considering all the items, the results are less consistent when only the items identified as showing differential difficulty are considered. Generally, 6 to 9 percent of the items were identified as being differentially difficult per comparison, with a greater number of items flagged in the female/male analyses than in the Black/White or Hispanic/White analyses. There was little overlap among the items flagged in the three different analyses. Although analyses suggested some content characteristics that may be related to differential item functioning, the findings should be considered exploratory due to the small number of items studied and the post hoc nature of the sample. Also, the findings should not be assumed to apply to populations other than NTE test takers. (113pp.)

Read More