On level and powee of anderson and hauck's peocedure for testing equivalence in compative bioavailability
- 1 January 1987
- journal article
- research article
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 16 (9) , 2771-2778
- https://doi.org/10.1080/03610928708829538
Abstract
The paper refers to a new procedure of Anderson and Hauck for the well-known problem of deciding whether two means of normal distributions with common but unknown variance differ within an admissible range. Analytical expressions for level and power of this method are derived and some numerical examples are given. A bound for the difference between real and no¬minal level is obtained which does not involve the variance.Keywords
This publication has 3 references indexed in Scilit:
- A new procedure for testing equivalence in comparative bioavailability and other clinical trialsCommunications in Statistics - Theory and Methods, 1983
- Statistical analysis of bioavailability studies: Parametric and nonparametric confidence intervalsEuropean Journal of Clinical Pharmacology, 1983
- Errors in Normal Approximations to the $t,\tau,$ and Similar Types of DistributionThe Annals of Mathematical Statistics, 1956