It is well-known that in a given decision problem if T is a sufficient statistic then given any decision function there exists an equivalent decision function that depends on the sample point only through T. In this note, it is shown that the same result does not hold if conditional inference is used unless the sufficient statistic satisfies an additional condition.
|Original language||English (US)|
|Number of pages||3|
|Journal||Statistics and Probability Letters|
|State||Published - Jul 13 1993|
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty