Mathematics course placement procedures have received significant scholarly attention in recent years. However, this research largely focuses on placement into community college development education coursework in a context where course recommendations are binding. In contrast, this paper examines non-binding mathematics placement procedures at a large flagship four-year university. We use support vector machines and a regression discontinuity design to study the causal effect of following mathematics placement recommendations on both short- and long-term outcomes. Our results show significant heterogeneity along gender and racial identities for which students follow their recommendation. Moreover, the decision to take calculus could potentially have negative effects on students’ outcomes. Our results offer insight for student placement policies at four-year universities.