High-dimensional Full-information Item Factor Analysis
Item factor analysis is unexcelled as a technique of exploration and discovery in the study of behavior. From binary-scored responses to a multiple-item test or scale, it determines the dimensionality of individual variation among the respondents and reveals attributes of the items defining each dimension. In practical test construction and scoring, it is the best guide to item selection and revision, as well as an essential preliminary step in justifying unidimensional IRT item analysis. Because hundreds of items may be analyzed jointly, the detail and generality that may be achieved exceeds that of any other procedure for exploring relationships among responses.
KeywordsMarginal Probability Item Parameter Quadrature Point Hadamard Matrix Likelihood Equation
Unable to display preview. Download preview PDF.
- Bock, R. D. (1985 reprint).Multivariate statistical methods in behavioral research.Chicago: Scientific Software International.Google Scholar
- Bock, R. D. (1989). Measurement of human variation: A two-stage model. In R. D. Bock (Ed.)Multilevel analysis of educational data (pp.319–342). New York: Academic Press.Google Scholar
- Connor, W. S. & Zelen, M. (1956).Fractional factorial experimental designs for factors at three levels.Washington, DC: U. S. Bureau of Standards.Google Scholar
- Householder, A. S. (1964).The theory of matrices in numerical analysis.New York: Blaisdell.Google Scholar
- Schilling, S. G. (1993). Advances in full-information factor analysis using the Gibbs sampler. University of Chicago, Department of Psychology. Unpublished dissertation.Google Scholar