Background: Is it possible to reach performance equality between boys and girls in a science class? Given the stereotypes targeting their groups in scientific domains, diagnostic contexts generally lower girls' performance and non-diagnostic contexts may harm boys' performance. Aim: The present study tested the effectiveness of a mastery-oriented assessment, allowing both boys and girls to perform at an optimal level in a science class. Sample: Participants were 120 boys and 72 girls (all high-school students). Methods: Participants attended a science lesson while expecting a performance-oriented assessment (i.e., an assessment designed to compare and select students), a mastery-oriented assessment (i.e., an assessment designed to help students in their learning), or no assessment of this lesson. Results: One way to increase girls' performance on a science test without harming boys' performance is to present assessment as a tool for improving mastery rather than as a tool for comparing performances. Conclusions: One way to increase girls' performance on a science test without harming boys' performance is to present assessment as a tool for improving mastery rather than as a tool for comparing performances.