Current quantitative gender and mathematics research exists as separate ability and postsecondary access studies, each drawing different conclusions. But if gender, achievement and enrollment (measurable proxies) are modeled together it will be possible to investigate possible associations. Using 106,473 standardized Grades 3, 6 and 9 Ontario student test responses from a single cohort, likelihood of enrolling in Grade 9 academic mathematics was modeled against elementary achievement and gender in a binary response model. Males achieved higher, (0.028=d=0.118), and occupied both achievement extremes in greater numbers, (1.04=VR=1.10), while females were 1.5 times more likely to enroll in academic courses. These paradoxical results are discussed in relation to the utility of achievement and enrollment as research metrics in gender and mathematics research.