You are not currently logged in.
Access JSTOR through your library or other institution:
If You Use a Screen ReaderThis content is available through Read Online (Free) program, which relies on page scans. Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
On Methods for Comparing Contingency Tables
Leo A. Goodman
Journal of the Royal Statistical Society. Series A (General)
Vol. 126, No. 1 (1963), pp. 94-108
Stable URL: http://www.jstor.org/stable/2982447
Page Count: 15
You can always find the topics here!Topics: Degrees of freedom, Null hypothesis, Standard error, Consistent estimators, Statistical discrepancies, Statistics, Statism, Statistical estimation, Standard deviation, Logical givens
Were these topics helpful?See somethings inaccurate? Let us know!
Select the topics that are inaccurate.
Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
Preview not available
In this paper we show that the standard error formulae given by Lewis (1962) and others for the Stouffer method of comparison of sets of 2 × 2 tables do not lead, in general, to consistent estimators of the appropriate standard deviations. Simple modifications of these formulae, which do lead to consistent estimators, are presented here. Since the relative difference between the standard errors published earlier and the consistent standard errors can be large, the earlier formulae should be replaced by the modified ones. For the analysis of K 2 × 2 tables (K ⩾ 2), we present here generalizations of the Stouffer method of comparison and also generalizations of the methods of analysis discussed by Berger (1961) for the special case where K = 2. A simple test of zero second-order interaction in a K × 2 × 2 contingency table is presented here. We also present some methods for analysing K I × J tables (describing attitude changes), and some methods of analysis suitable for the situation where the response categories (i.e. the row and column categories in the tables) are ordinally scaled.
Journal of the Royal Statistical Society. Series A (General) © 1963 Royal Statistical Society