The conditional test is compared with other relevant test procedures for testing parameters of Normal and Exponential populations. It is seen that under certain conditions the conditional test is more powerful than the relevant unconditional tests. An example is provided where the conditional test given the ancillary is uniformaly more powerful than the obvious unconditional test.
@article{urn:eudml:doc:40709,
title = {Some aspects of the conditional inference when the sample size is random.},
journal = {Trabajos de Estad\'\i stica e Investigaci\'on Operativa},
volume = {34},
year = {1983},
pages = {121-130},
zbl = {0729.62506},
language = {en},
url = {http://dml.mathdoc.fr/item/urn:eudml:doc:40709}
}
Swain, A.K.P.C. Some aspects of the conditional inference when the sample size is random.. Trabajos de Estadística e Investigación Operativa, Tome 34 (1983) pp. 121-130. http://gdmltest.u-ga.fr/item/urn:eudml:doc:40709/