Randomization and Significance Testing in Clinical Trials

Abstract: Randomization at the beginning and significance testing at the end of a clinical trial are inextricablv related. The evaluation of a randomized trial requires a randomization test, a random sample test would be misplaced. The verification of the success of the randomization by significance testing is not appropriate. The outcomes of clinical trials hardly permit generalization; the concept of randomized studies does not guarantee representative results. The investigation of a cause-effect relation requires an experimental approach which entirely makes use of a random process.

Keyword(s): Klinische Therapieprüfung

Zurück | Weiter