Exact Conditional Tests for Hierarchical Models in Multidimensional Contingency Tables
- Resource Type
- research-article
- Authors
- Morgan, W. Meade; Blumenstein, Brent A.
- Source
- Journal of the Royal Statistical Society. Series C (Applied Statistics), 1991 Jan 01. 40(3), 435-442.
- Subject
- Conditional Tests
Exact Tests
Hierarchical Models
Multidimensional Contingency Tables
Algorithms
Significance level
Subroutines
Statistics
Computer software
Sample size
Ratios
Parametric models
Hemorrhage
Asymptotic methods
- Language
- English
- ISSN
- 00359254
14679876
Exact conditional inference is routinely used for contingency tables with restricted dimensions such as 2× 2, I× J and 2× 2× K tables. This paper introduces an algorithm and computer subroutine for the exact conditional test of hierarchical models in multidimensional contingency tables. The algorithm depends on complete enumeration, and the feasibility of using this algorithm is illustrated. Exact conditional inference is useful for sparse multidimensional tables or multidimensional tables with small sample size.