Abstract: International students are often well represented in graduate programmes in North America and Europe. Information on foreign countries' education systems and grading schemes is available but cross-country comparisons are often challenging and highly subjective. Therefore, universities have a clear need for calibrating admissions of international students to ensure a fair and cost effective selection process. By comparing the performance of international students in their host institution with their entry qualifications we devise a simple approach to detecting systematic biases in the perceived quality of the applicants and propose corrective actions. We find that by using public information on cross-country comparisons of academic qualifications, country selection biases can occur and produce a substantial impact on international students' performance and failure rates. Our model is based on admission data that are routinely collected by universities which should ensure its broad applicability.