In this study, we investigated undergraduate mathematics students' (N = 267) attitudes towards proving. The students were taking an introduction-to-proof type course that was situated at the beginning of the mathematics curriculum and lasted for one term. Four attitude variables were measured at the beginning and at the end of the course with a new self-report instrument: self-efficacy, anxiety, appreciation and motivation. The instrument was based on two existing instruments on mathematics attitudes and proof-related self-efficacy. We studied how these four attitude variables were related to the students' prior skills and their gender at the beginning of the course, how the attitude variables changed during the course, and how they affected the students' performance in the final course project. Our results indicate that students' prior performance is linked to their proof-related self-efficacy, anxiety and motivation at the beginning of the course. Female students exhibited lower efficacy and motivation levels than male students. During the focus course, students' self-efficacy increased and anxiety decreased. The gender gaps in self-efficacy and motivation persisted throughout the course. In addition, high motivation in the beginning of the course predicted good performance in the final project. Based on our results, we conclude that an introductory course on proving can enhance students' attitudes, and we suggest that these attitudes are taken into account in teaching as they can affect students' performance. Finally, we urge researchers and professionals to earnestly consider ways to mitigate gender differences in mathematics.