ERIC Number: ED469371
Record Type: Non-Journal
Publication Date: 1999-Mar
Reference Count: N/A
An Overview of Nonlinear Factor Analysis and Its Relationship to Item Response Theory. Statistical Report. LSAC Research Report Series.
De Champlain, Andre F.
This paper provides a brief overview of some of the research that has examined the relationship between item response theory (IRT) and nonlinear factor analysis (NFA), and outlines three NLFA models, emphasizing their major strengths and weaknesses for practical applications. The focus is on: (1) R. McDonald's polynomial approximation of a normal ogive model (1967, 1982); (2) A. Christoffersson's (1975) and B. Muthen's (1984) factor analytic model for dichotomous variables; and (3) the full information factor analytic model (R. Bock and M. Aitken, 1981; R. Bock, R. Gibbons, and E. Muraki, 1988). Items from two Law School Admission Test (LSAT) forms were calibrated using these three models to assess the degree of comparability of the IRT parameter estimates using these procedures. Findings show that the parameter estimates tended to be quite similar, regardless of the calibration procedure. Additional research should be conducted with a larger number of LSAT forms before reaching any conclusions about the comparability of the calibration procedures. (Contains 2 tables, 2 figures, and 55 references.) (Author/SLD)
Publication Type: Information Analyses; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers - Assessments and Surveys: Law School Admission Test