Goodman and Kruskal (1972) developed the asymptotic sampling theory for the association measure Tau b. Their formula has been modified by Bishop, Fienberg, and Holland (1978) and generalized for the partial analog by Gray and Williams (1981). Unfortunately, both modifications are incorrect, which is shown in this article. Also, the correct formula for the variance of the partial Tau b is given.
Get full access to this article
View all access options for this article.
References
1.
BISHOP, M. M.
, S. E. FIENBERG, and P. W. HOLLAND (1978) Discrete Multivariate Analysis. Theory and Practice. Cambridge: MIT Press.
2.
CRAMÉR, H.
(1946) Mathematical Methods of Statistics. Princeton, NJ: Princeton Univ. Press.
3.
GOODMAN, L. A.
and W. H. KRUSKAL (1954) “Measures of association for cross-classifications.”J. Amer. Stat. Assn.49: 732-764.
4.
GOODMAN, L. A.
and W. H. KRUSKAL(1963) “Measures of association for cross-classifications III: approximate sampling theory.”J. Amer. Stat. Assn.58: 310-364.
5.
GOODMAN, L. A.
and W. H. KRUSKAL(1972) “Measures of association for cross-classifications IV: simplification of asymptotic variances.”J. Amer. Stat. Assn.67: 415-421.
6.
GRAY, L. N.
and J. S. WILLIAMS (1981) “Goodman and Kruskal's Tau b. Multiple and partial analogs.”Soc. Methods and Research10: 50-62.
7.
ZIMMERMANN, K. F.
and A. ZIMMERMANN-TRAPP (1982) “Arbeitsmarktorientierung und Weiterbildung. Eine informationsökonomische Analyse des Weiterbildungsverhaltens von Frauen unter Verwendung log-linearer Wahrscheinlichkeitsmodelle,” in H. König (ed.) Ausbildung und Arbeitsmarkt. Frankfurt, Germany: Lang-Verlag.
