There are several reasons to be cautious about accepting the claims contained in Allen's (1991) meta‐analysis of the research literature concerning the relative persuasive effectiveness of one‐sided and two‐sided messages. The classification of the message types involved in a number of studies is questionable, and the literature search underlying the review appears to be incomplete. These problems indicate that it would be imprudent to accept the review's conclusions at face value, but, more important, they also underscore the importance of evolving standards for the assessment of meta‐analytic research.