Exact inference for categorical data: recent advances and continuing controversies
- 14 August 2001
- journal article
- research article
- Published by Wiley in Statistics in Medicine
- Vol. 20 (17-18) , 2709-2722
- https://doi.org/10.1002/sim.738
Abstract
Methods for exact small‐sample analyses with categorical data have been increasingly well developed in recent years. A variety of exact methods exist, primarily using the approach that eliminates unknown parameters by conditioning on their sufficient statistics. In addition, a variety of algorithms now exist for implementing the methods. This paper briefly summarizes the exact approaches and describes recent developments. Controversy continues about the appropriateness of some exact methods, primarily relating to their conservative nature because of discreteness. This issue is examined for two simple problems in which discreteness can be severe – interval estimation of a proportion and the odds ratio. In general, adjusted exact methods based on the mid‐P‐value seem a reasonable way of reducing the severity of this problem. Copyright © 2001 John Wiley & Sons, Ltd.Keywords
This publication has 101 references indexed in Scilit:
- Approximate Is Better than "Exact" for Interval Estimation of Binomial ProportionsThe American Statistician, 1998
- Approximate Conditional Inference in Exponential Families via the Gibbs SamplerJournal of the American Statistical Association, 1994
- Methods for Exact Goodness-of-Fit TestsJournal of the American Statistical Association, 1992
- An Evaluation of Some Tests of Trend in Contingency TablesJournal of the American Statistical Association, 1992
- Computing Exact Distributions for Polytomous Response DataJournal of the American Statistical Association, 1992
- On the Logical Justification of Conditional Tests for Two-By-Two Contingency TablesThe American Statistician, 1991
- Importance Sampling for Estimating Exact Probabilities in Permutational InferenceJournal of the American Statistical Association, 1988
- A Warning on the Use of Chi-Squared Statistics with Frequency Tables with Small Expected Cell CountsJournal of the American Statistical Association, 1988
- Binomial Confidence IntervalsJournal of the American Statistical Association, 1983
- Significance Tests in Discrete DistributionsJournal of the American Statistical Association, 1961