An important part of evidence-based practice is to include client preferences in the treatment decision-making process. However, based on previous reviews of the literature there is some question as to whether including client preferences actually has an effect on treatment outcome. This meta-analytic review summarized data from over 2,300 clients across 26 studies comparing the treatment outcome differences between clients matched to a preferred treatment and clients not matched to a preferred treatment. The findings indicate a small significant effect (r=.15, CI.95: .09 to .21) in favor of clients who received a preferred treatment. The binomial effect size indicated that matched clients have a 58% chance of showing greater improvement, and further analysis indicate that they are about half as likely to drop-out of treatment when compared with clients not receiving a preferred treatment. Study design was seen to be a moderating variable in that partially randomized preference trials may underestimate the treatment preference effect. Implications for best practice standards are discussed. © 2009 Wiley Periodicals, Inc. J Clin Psychol 65:1–14, 2009.