Date of this Version
Statistical testing of null hypotheses recently has come under fire in wildlife sciences (Cherry 1998; Johnson 1999; Anderson et al. 2000, 2001). In response to this criticism, Robinson and Wainer (2002) provide some further background information on significance testing; they argue that significance testing in fact is useful in certain situations. I counter by suggesting that such situations rarely arise in our field. I agree with Robinson and Wainer that replication is the key to scientific advancement. I believe, however, that significance testing and resulting P-values frequently are confused with issues of replication. Any single study can yield a P-value, but only consistent results from truly replicated studies will advance our understanding of the natural world.