This study investigates the impact of single-sex versus coeducational schooling on students' decisions to pursue STEM fields at the university level. Using administrative data from eight undergraduate cohorts (2012-2019) at a prominent Australian university, we compare students with similar Australian Tertiary Admissions Ranks (ATARs) who could have feasibly enrolled in either school type of comparable quality under different circumstances. We control for individual characteristics and the academic quality of the high schools attended. Our primary outcomes are the proportion of students from each school type choosing a STEM major and their weighted average marks for each year of university studies. Contrary to expectations, we find no evidence that a single-sex high school background increases STEM participation among girls at the university level. Interestingly, students from single-sex high schools show a higher propensity to choose a business major. Additionally, we find that the linear correlation between ATAR scores and first-year university grades is approximately 0.4. However, our analysis suggests that this relationship is better characterized as nonlinear rather than linear.