This paper looks at horizontal sex segregation in education as a factor contributing to gender segregation in the labor market. Economic theories fail to explain why women with the same years of schooling and educational attainment as men are under-represented in many technical degrees, which typically lead to better paid occupations. Following Akerlof and Kranton (2000), I research whether gender identity affects boys' and girls' educational choices and when the gendered pattern appears first. Further, I test the hypothesis that single-sex schools attenuate the influence of gender-stereotypes. I use the National Pupil Database, which is a register of all pupils enrolled in state maintained schools in England and I focus on students in lower and upper secondary education. Results from my analysis suggest that gender stereotyping affects educational choices from the age of 14 and this effect is larger for girls than for boys. I also find that attending a sixth-form-single-sex school leads students to a less stereotyped educational choice, after controlling for endogenous self-selection into single-sex schools. This suggests that gender preferences can be modified by the environment.