Social-psychological interventions in education have used a variety of "self-persuasion" or "saying-is-believing" techniques to encourage students to articulate key intervention messages. These techniques are used in combination with more overt strategies, such as the direct communication of messages in order to promote attitude change. However, these different strategies have rarely been systematically compared, particularly in controlled laboratory settings. We focus on 1 intervention based in expectancy-value theory designed to promote perceptions of utility value in the classroom and test different intervention techniques to promote interest and performance. Across 3 laboratory studies, we used a mental math learning paradigm in which we varied whether students wrote about utility value for themselves or received different forms of directly communicated information about the utility value of a novel mental math technique. In Study 1, we examined the difference between directly communicated and self-generated utility-value information and found that directly communicated utility-value information undermined performance and interest for individuals who lacked confidence, but that self-generated utility had positive effects. However, Study 2 suggests that these negative effects of directly communicated utility value can be ameliorated when participants are also given the chance to generate their own examples of utility value, revealing a synergistic effect of directly communicated and self-generated utility value. In Study 3, we found that individuals who lacked confidence benefited more when everyday examples of utility value were communicated, rather than career and school examples.