Gender differences in participation and performance at "high stakes" examinations have received much public attention, which has often focused on mathematics and science subjects. This paper describes the innovative forms of assessment introduced into mathematics and science subjects within the Victorian Certificate of Education (VCE) system. Results from these subjects are examined for patterns of gender differences in participation and performance over the period 1994-1999. A larger proportion of males than females studied all the VCE science and mathematics subjects except Biology and Psychology over this period. Based on study scores, females, on average, out-performed males in almost all VCE science and mathematics subjects in nearly every year from 1994-1999. As exceptions to the patterns, males out-performed females in Chemistry and Mathematical Methods. Results from a general ability test are used to question the legitimacy of gender comparisons in subjects in which enrolment is no longer compulsory. The data do not support simplistic conclusions about gender differences in participation and performance. (Contains 5 tables and 5 figures.)