NotesFAQContact Us
Search Tips
ERIC Number: ED377227
Record Type: Non-Journal
Publication Date: 1993-Apr
Pages: 27
Abstractor: N/A
Reference Count: N/A
Differential Item Functioning vs Differential Test Functioning.
Bergstrom, Betty A.; And Others
A problem that arises when a differential item functioning (DIF) study is done with samples of examinees differing in ability is examined. A test may function differently when the populations from which the items are calibrated are not of equal ability. Since the lower ability examinees get many difficult items incorrect, the spread (standard deviation) of item calibrations may differ. The difference in how the test functions, in terms of item spread, must be addressed before differential item function can be explored. Several methods to account for differences in standard deviation are reported. The problem is studied using data from the Michigan Educational Assessment Program (MEAP) for 619 eleventh and twelfth graders from the 1991 administration of the MEAP mathematics test. Either linear transformation or subsampling helps eliminate differences in standard deviation, but subsampling is preferable when there is a large enough sample. Five figures and three tables illustrate the analysis, along with eight figures in an appendix describing sample groups. (Contains 9 references.) (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A