And Yet the Small-Sample Effect Does Hold: Reply to Juslin and Olsson (2005) and Anderson, Doherty, Berg, and Friedrich (2005).