Vol: 32 Page: 321

Authors:
Jolanta Grala-Michalak
Krystyna Katulska

Title:
 A note on power of ML tests for normally and student's t distributed data

Language: English

Keywords:
likelihood ratio test
multivariate Student's distribution
completely randomized block design

Summary:
This paper deals with comparison of the power of two likelihood ratio tests in the Gauss-Markoff model for the completely randomized block design. In one of the tests we assume that the vector of errors follows Student’s t distribution, whereas in another is normally distributed.