Replications and t-Tests


Scott Armstrong adds to the thread

 ARC: Community: ELMAR: Posting

areas: methods: dialog

Replication Reply to John Rossiter and the Silent Majority

John Rossiter’s position that statistical significance is related to replication is commonly held among researchers. For example, a 1986 survey of researchers by M. L. Oakes found that 60% of research psychologists and advanced graduate students associate t tests with replication.

So who am I to argue with commonly held beliefs?

On the other hand, Schmidt (1996, p. 124-125) summarized Oakes’ survey and stated that it has long been known that many researchers cling to this false belief. He also explains why it is false. I found the paper by Schmidt to be convincing.


Schmidt, Frank L. (1996), "Statistical Significance Testing and Cumulative Knowledge in Psychology: Implications for Training of Researchers," Psychological Methods, 1, 115-129.

Scott Armstrong