radio, or print); and the use of supplementary components, such as group discussion, role play, or supportive services. The authors acknowledge significant methodological challenges for both the research evaluating media interventions and the meta-analysis, and the effect sizes they found were small. However, they conclude that media interventions can be effective, and that the wide reach of such interventions can potentially translate a small effect into significant cumulative changes for large numbers of young people.
The lifetime prevalence of eating disorders, including anorexia nervosa, bulimia nervosa, and binge eating disorder, is relatively small, more common among females, and most likely to occur during the teen years (Stice and Peterson, 2007). In a meta-analysis of 53 randomized and quasi-experimental trials focused on prevention of eating disorders, Stice and Shaw (2004) found, on average, significant effects (generally small to modest) for each of the included dependent variables: body mass, thin-ideal internalization, body dissatisfaction, dieting, negative affect, and eating pathology. Some effects were detectable as much as two years after the intervention. The effect sizes were smaller for universal interventions, which included many participants not at risk for eating disorders. Didactic programs were less effective than those that engaged participants in interactions. Single-session programs were less effective than longer ones, and programs were more effective if they targeted those over age 15. Interventions that simply provided education about eating disorders were significantly less effective than other interventions for most outcomes. The effective interventions varied in content and included ones that focused on resistance to cultural pressure for thinness, addressed body dissatisfaction, and taught healthy weight management. A meta-analysis of five studies of Internet-based interventions to prevent eating disorders found no statistical significance for pooled outcome data but recommended additional research given the small number of studies (Newton and Ciliska, 2006). Stice and Shaw (2004) similarly point to the need for improved methodological rigor and theoretical rationale in order to progress from promising to conclusive interventions.
There has been limited work on early prevention of psychotic disorders. Given the severity of such disorders as schizophrenia and bipolar disorder (McFarlane, 2007) and their extraordinarily high associated lifetime risk for suicide (Palmer, Pankratz, and Bostwick, 2005) and early mortality (Fenton,