Accession Number : AD0762110

Title :   Multivariate Test Procedures for Unknown Means When Sampling from Bivariate Normal Populations.

Descriptive Note : Technical note,

Corporate Author : ARMY MISSILE COMMAND REDSTONE ARSENAL ALA SYSTEMS ANALYSIS DIV

Personal Author(s) : Colvin,Charles E.

Report Date : MAR 1973

Pagination or Media Count : 23

Abstract : AL TESTS), MATHEMATICAL PREDICTION, PROBABILITY DENSITY FUNCTIONS, MATRICES(MATHEMATICS)NORMAL DENSITY FUNCTIONS, BIVARIATE ANALYSIS, STATISTICAL INFERENCEThe report investigates and presents statistical testing procedures when sampling from bivariate normal populations for which inferences are drawn concerning unknown means. Multivariate hypothesis testing procedures for testing for unknown means in the bivariate case are presented. Pitfalls and shortcomings of using separate hypotheses for each unknown mean are discussed. An example for the bivariate case is shown whereby two different conclusions can be reached using the same experimental data. One set of conclusions is made using the multivariate procedure (correct procedure), and a completely different set of conclusions is made using the classical hypothesis testing procedure for each unknown mean. (Author)

Descriptors :   (*MULTIVARIATE ANALYSIS, STATISTICAL TESTS), (*SAMPLING, STATISTICAL TESTS), MATHEMATICAL PREDICTION, PROBABILITY DENSITY FUNCTIONS, MATRICES(MATHEMATICS)

Subject Categories : Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE