Estimation of the Reliability of Skill Tests

Abstract This article seeks to clarify the concept of test reliability and to differentiate between the theoretical definition and the methods of estimation. In the first part, the split-halves and test-retest methods are critically examined in the context of a typical skills test. The broader and more comprehensive definition of error under the test-retest method is noted. In the second part, the use of analysis of variance techniques in reliability studies is illustrated by application to a badminton wall volley test. The advantages of this approach over traditional approaches is discussed and possible applications cited.