Citation

Duncan OD. Soc. Sci. Res. 1983; 12(4): 393-400.

Copyright

(Copyright © 1983, Elsevier Publishing)

DOI

10.1016/0049-089X(83)90026-1

PMID

unavailable

Abstract

Kempf's model was used by W. F. Kempf and P. Hampapa (1977, "Conditional inference for the dynamic test model," in Mathematical Models for Social Psychology (W. F. Kempf and B. H. Repp, Eds.), Chap. 3, pp. 81-100, Wiley, New York) to analyze the cross-classification of responses to four items on sex typing of children's household chores from the 1971 Detroit Area Study. The model alters the simple Rasch measurement model for dichotomous items to include transfer effects, so that responses are allowed to be serially dependent (G. Rasch, 1966, "An individualistic approach to item analysis," in Readings in Mathematical Social Science (P. F. Lazarsfeld and N. W. Henry, Eds.), pp. 89-107, MIT Press, Cambridge, 1980; Probabilistic Models for Some Intelligence and Attainment Tests, expanded ed., Univ. of Chicago Press, Chicago). Although Kempf and Hampapa did not find strong reason to reject the dynamic model, the present analysis indicates that it should be rejected in favor of a two-dimensional Rasch model.