2005•Encyclopedia of BiostatisticsRequires access
Hypothesis Testing
David S. Salsburg
Open publisher page 0 citations
Abstract
Abstract A significance test of a null hypothesis involves the choice of a test statistic whose null distribution is known, and a judgment as to whether, or how well, the observed value of the statistic accords with the known distribution. The article reviews the history of this approach, with emphasis on the Neyman–Pearson formulation, taking account of the behavior of the test statistic under alternative hypotheses.