To enhance the prejudice-reducing effects of imagined contact we investigated a novel form of imagined contact ("balanced similarity") which emphasized both similarities and differences between the ingroup(er) and the outgroup(er). Experiment 1 compared balanced similarity with conditions inducing only differences or only similarities. "Balanced similarity" led to more positive outgroup attitudes; its differences with the "high similarity" condition were mediated by reduced distinctiveness threat, whereas its differences with the "low" similarity condition were mediated by higher perceived intergroup similarity. Experiment 2 compared the "balanced similarity" imagined contact scenario with the "standard" (positive) imagined contact scenario (Crisp, Stathi, Turner, & Husnu, 2008), and found that both conditions promoted equally favourable attitudes that were significantly more positive than in the control condition. However, only the "balanced similarity" imagined contact condition differed from the control condition on intergroup anxiety and contact self-efficacy. The "balanced similarity" condition also had an indirect effect (via self-efficacy) on positive action tendencies towards the outgroup. We discuss the utility of "balanced similarity" imagined contact, especially where contact is limited and conflict is present.