In Turkey, as in many other countries, female students perform better in high school and have higher test scores than males. Nevertheless, men still predominate at highly selective programs that lead to high-paying careers. The gender gap at elite schools is particularly puzzling because college admissions are based entirely on nationwide exam scores. Using detailed administrative data from the centralized college entrance system, I study the impact of gender differences in preferences for programs and schools on the allocation of students to colleges. Controlling for test score and high school attended, I find that females are more likely to apply to lower-ranking schools, whereas males set a higher bar, revealing a higher option value for re-taking the test and applying again next year. I also find that females and males value program attributes differently, with females placing more weight on the distance from home to college, and males placing more weight on program attributes that are likely to lead to better job placements. Together, these differences in willingness to be unassigned and in relative preferences for school attributes can explain much of the gender gap at the most elite programs.