ERIC Number: ED333050
Record Type: RIE
Publication Date: 1991-Apr
Reference Count: N/A
Effects of Linking Methods on Detection of DIF.
Kim, Seock-Ho; Cohen, Allan S.
Studies of differential item functioning (DIF) under item response theory require that item parameter estimates be placed on the same metric before comparisons can be made. Evidence that methods for linking metrics may be influenced by the presence of differentially functioning items has been inconsistent. The effects of three methods for linking metrics on detection of DIF were studied. The methods included: (1) a weighted mean and sigma method; (2) the test characteristic curve method; and (3) the minimum chi-square method. Both iterative and non-iterative linking procedures were compared for each method. A two-parameter logistic item characteristic model was used to generate eight simulated data sets. A 60-item test and sample sizes of 300 and 600 were generated for each data set. Results indicate that detection of DIF following linking via the test characteristic curve method gave the most accurate results when the sample size was small. When the sample size was large, results for the three methods were essentially the same. Iterative linking provided a substantial improvement in detection of DIF over non-iterative linking. A 28-item list of references is included. Seven tables present study findings, and an appendix summarizes the three methods. (Author/SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Item Characteristic Function; Item Parameters; Linking Metrics
Note: Paper presented at the Annual Meeting of the National Council on Measurement in Education (Chicago, IL, April 4-6, 1991).