Existing research into students' preferences for assessment methods has been developed from a restricted sample: in particular, the voice of students in the 'hard-pure sciences' has rarely been heard. We conducted a mixed method study to explore mathematics students' preferences of assessment methods. In contrast to the message from the general assessment literature, we found that mathematics students differentially prefer traditional assessment methods such as closed book examination; they perceive them to be fairer than innovative methods and they perceive traditional methods to be the best discriminators of mathematical ability. We also found that although students prefer to be assessed by traditional methods they are also concerned by the mix of methods they encounter during their degree, suggesting that more account needs to be taken about the students' views of this mix. We discuss the impact of the results on the way general findings about assessment preference should be interpreted.