Abstract
In this paper algorithms are described for obtaining the maximum likelihood estimates of the parameters in loglinear models. Modified versions of the iterative proportional fitting and Newton-Raphson algorithms are described that work on the minimal sufficient statistics rather than on the usual counts in the full contingency table. This is desirable if the contingency table becomes too large to store. Special attention is given to loglinear IRT models that are used for the analysis of educational and psychological test data. To calculate the necessary expected sufficient statistics and other marginal sums of the table, a method is described that avoids summing large numbers of elementary cell frequencies by writing them out in terms of multiplicative model parameters and applying the distributive law of multiplication over summation. These algorithms are used in the computer program LOGIMO. The modified algorithms are illustrated with simulated data.
Similar content being viewed by others
References
Adby, P. R., & Dempster, M. A. H. (1974).Introduction to optimization methods. London: Chapman and Hall.
Agresti, A. (1984).Analysis of ordinal categorical data. New York: Wiley.
Andersen, E. B. (1972). The numerical solution of a set of conditional estimation equations.Journal of the Royal Statistical Society, Series B, 34, 42–54.
Andersen, E. B. (1973). Conditional inference and multiple choice questionnaires.British Journal of Mathematical and Statistical Psychology, 26, 31–44.
Andersen, E. B. (1980).Discrete statistical models with social science applications. Amsterdam: North Holland.
Baker, R. J., & Nelder, J. A. (1978).The GLIM system: Generalized linear interactive modeling. Oxford: The Numerical Algorithms Group.
Bock, R. D. (1975).Multivariate statistical methods in behavioral research. New York: McGraw-Hill.
Cressie, N., & Holland, P. W. (1983). Characterizing the manifest probabilities of latent trait models.Psychometrika, 48, 129–142.
Deming, W. E., & Stephan, F. F. (1940). On a least squares adjustment of a sampled frequency table when the expected marginal totals are known.Annals of Mathematical Statistics, 11, 427–444.
Dongarra, J. J., Bunch, J. R. Moler, C. B., and Stewart, G. W. (1979).LINPACK user's guide. Philadelphia, PA: SIAM.
Duncan, O. D. (1984). Rasch measurement: Further examples and discussion. In C. F. Tuner & E. Martin (Eds.),Surveying subjective phenomena, Vol. 2 (pp. 367–403). New York: Russell Sage Foundation.
Duncan, O. D., & Stenbeck, M. (1987). Are likert scales unidimensional?Social Science Research, 16, 245–259.
Fischer, G. H. (1974).Einführung in die Theorie psychologischer Tests [Introduction to the theory of psychological tests]. Bern: Huber. (In German)
Formann, A. K. (1986). A note on the computation of the second-order derivatives of the elementary symmetric functions in Rasch models.Psychometrika, 51, 335–339.
Gill, P. E., Murray, W., & Wright, M. H. (1991).Numerical linear algebra and optimization. Redwood City: Addison Wesley.
Glas, C. A. W., (1989).Estimation and testing Rasch models. Unpublished doctoral dissertation, University of Twente, Enschede, The Netherlands. (ISBN 90-9003078-6)
Glas, C. A. W. (1989).Contributions to estimating and testing Rasch models. Arnhem, The Netherlands: Cito.
Goodman, L. A., & Fay, R. (1974).ECTA program, description for users. Chicago: University of Chicago, Department of Statistics.
Gustafsson, J. E. (1977).The Rasch model for dichotomous items: Theory, applications and a computer program (Reports form the Institute of Education, Rep. No. 85). Göteborg: University of Göteborg.
Gustafsson, J. E. (1980). A solution of the conditional estimation problem for long tests in the Rasch model for dichotomous items.Educational and Psychological Measurement, 40, 377–385.
Haberman, S. J. (1979).Analysis of qualitative data: New developments, Vol. 2. New York: Academic Press.
Holland, P. W., & Thayer, D. T. (1987).Notes on the use of log-linear models for fitting discrete probability distributions (Technical Report No. 87-79). Princeton, NJ: Educational Testing Service.
Kelderman, H. (1984). Loglinear Rasch model tests.Psychometrika, 49, 223–245.
Kelderman, H. (1989a). Item bias detection using loglinear IRT.Psychometrika, 54, 681–698.
Kelderman, H. (1989b, March).Loglinear multidimensional IRT models for polytomously scored items. Paper read at the Fifth International Objective Measurement Workshop, Berkeley. (ERIC document Reproduction Service No. ED 308 238)
Kelderman, H. (1991, April).Estimating and testing a multidimensional Rasch model for partial credit scoringg. Paper read at the Annual Meeting of the American Educational Research Association, Chicago, IL. (In review)
Kelderman, H., & Steen, R. (1988).LOGIMO: Loglinear IRT Modeling [Program manual]. Enschede, The Netherlands: University of Twente.
Kelderman, H. & Steen, R. (1988).LOGIMO I; A program for loglinear item response theory modeling. Twente, The Netherlands: University of Twente, Department of Education.
Masters, G. N. (1982). A Rasch model for partial credit scoring.Psychometrika, 47, 149–174.
Rasch, G. (1960).Probabilistic models for some intelligence and attainment tests. Copenhagen: Paedagogiske Institut.
Rasch, G. (1961). On general laws and the meaning of measurement in psychology.Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, 5, 321–333.
Rasch, G. (1980.Probabilistic models for some intelligence and attainment tests. Chicago: The University of Chicago Press.
Scheiblechner, H. (1971).A simple algorithm for CML-parameter estimation in Rasch's probabilistic measurement model with two or more categories of answers (Research Bulletin Nr. 5/71). Vienna, Austria: University of Vienna, Psychological Institute.
SPSS. (1988).SPSS User's Guide (2nd ed.). Chicago, IL: Author.
Tjur, T. (1982). A connection between Rasch's item analysis model and a multiplicative Poisson model.Scandinavian Journal of Statistics, 9, 23–30.
Verhelst, N. D., Glas, C. A. W., & van der Sluis, A. (1984). Estimation problems in the Rasch model: The basic symmetric functions.Computational Statistics Quarterly, 1, 245–262.
Author information
Authors and Affiliations
Additional information
The author thanks Wim J. van der Linden, Gideon J. Mellenberh and Namburi S. Raju for their valuable comments and suggestions.
Rights and permissions
About this article
Cite this article
Kelderman, H. Computing maximum likelihood estimates of loglinear models from marginal sums with special attention to loglinear item response theory. Psychometrika 57, 437–450 (1992). https://doi.org/10.1007/BF02295431
Issue Date:
DOI: https://doi.org/10.1007/BF02295431