Testing for independence in a two-way table: New interpretations of the chi-square statistic (Q1076450): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Removed claim: author (P16): Item:Q163054 |
||
Property / author | |||
Property / author: Efron, Bradley / rank | |||
Revision as of 22:32, 9 February 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Testing for independence in a two-way table: New interpretations of the chi-square statistic |
scientific article |
Statements
Testing for independence in a two-way table: New interpretations of the chi-square statistic (English)
0 references
1985
0 references
This paper is concerned with \(\chi^ 2\)-tests for the independence in contingency tables. In particular, it deals with the problem how to analyse a contingency table if the \(\chi^ 2\)-test rejects independence. The authors propose to consider the single alternative, consisting of the uniform distribution. They connect the hypothesis of independence and the hypothesis of uniformity by means of a one-parameter exponential family, and consider the statistical analysis of the parameter of this family. The paper is followed by a number of discussions by other authors, most of them criticize the approach proposed in this paper.
0 references
chi-square test for independence
0 references
overdispersion
0 references
volume tests
0 references
random effects for exponential families
0 references
contingency tables
0 references
one-parameter exponential family
0 references