A TEST OF A SPECIAL CASE OF TYPICALITY IN LINEAR DISCRIMINANT-ANALYSIS

Let us consider two populations having p-variate normal distributions with a common variance-covariance matrix where the parameters are know n. Let X denote the vector of observations on an experimental unit who se origin is uncertain. A test is proposed to test the hypothesis that X is distributed according to one of the two distributions against th e hypothesis of a third normal distribution with a mean that is a line ar combination of the means of the two populations. Simulations were p erformed to check the level and the power of the test under various hy potheses. An example is presented.