Building on research on gender stereotypes and a parallel-constraint-satisfaction theory on impression formation, this project investigates the effects of gender stereotypical and counter-stereotypical visuals on voters’ evaluations of political candidates with two pre-registered experimental studies. Study 1 (N = 1,225) is a conceptual replication of an online experiment on the effect of visual communication of fictional U.S. candidates, testing main assumptions in the context of real-world German candidates on X (formerly Twitter). In contrast to the original study, we find that visuals reinforcing masculine as well as feminine stereotypes can be detrimental to men candidates. Evaluations of women candidates were not affected by (counter) gender stereotypical visual information. Study 2 (N = 1,058) repeats Study 1 outside of an election context. Findings differ from the first study. Moreover, no effects of visual communication are found when prior attitudes are controlled. The project highlights the importance of replicating experimental findings in different contexts.