You are not currently logged in.
Access your personal account or get JSTOR access 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.
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
When testing an affine hypothesis in an exponential family the `ideal' procedure is to calculate the exact similar test, or an approximation to this, based on the conditional distribution given the minimal sufficient statistic under the null hypothesis. By contrast to this there is a `primitive' approach in which the marginal distribution of a test statistic is considered and any nuisance parameter appearing in the test statistic is replaced by an estimate. We show here that when using standardized likelihood ratio statistics the `primitive' procedure is in fact an `ideal' procedure to order O(n-3/2). As an example we consider inference for the mean of a log normal distribution in detail.
Biometrika © 1986 Biometrika Trust