Compelling arguments for reform of statistical practices have been made in many disciplines, in some cases over several decades, but achieving reform has proved difficult. We discuss how reform has progressed--or not progressed--in psychology, medicine, and ecology and describe case studies of attempts by pioneering journal editors to change statistical practices. Lessons for those seeking reform in education include the need to recognize the importance of journal editors and of provision of articles, books, and software that give practical guidance to researchers wishing to use the recommended statistical techniques. Research is required on recommended techniques so that statistical practice can become evidence based. Also, improvement in statistical practice should be encouraged along with improvement in the way a discipline theorizes.