NotesFAQContact Us
Search Tips
ERIC Number: ED244986
Record Type: Non-Journal
Publication Date: 1984-Apr
Pages: 46
Abstractor: N/A
Reference Count: N/A
A Study of Item Bias for Attitudinal Measurement Using Maximum Likelihood Factor Analysis.
Mayberry, Paul W.
A technique for detecting item bias that is responsive to attitudinal measurement considerations is a maximum likelihood factor analysis procedure comparing multivariate factor structures across various subpopulations, often referred to as SIFASP. The SIFASP technique allows for factorial model comparisons in the testing of various hypotheses concerning the interitem covariance matrices. To demonstrate the utility of this technique, the Inventory of Personal Investment (IPI) was administered to three groups: academics, management, and college students. The basic hypothesis was that academics, management, and students with identical standing on any one of the seven latent incentives scales would have identical expected observed scores on the same scale, i.e., an identical factor structure would exist across the groups. The results showed that consistencies within the biased items can be identified, and that 13 of the IPI items did not provide equivalent measurement across the three subpopulations of this study. The maximum likelihood factor analysis procedure provided many advantages over the classical test theory and item response theory approaches for the detection of item bias in attitudinal instruments. (BW)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Audience: Researchers
Language: English
Sponsor: N/A
Authoring Institution: N/A