In this project I look at the results of Polish matriculation exams (matura) from the gender perspective. Specifically, I investigate if male and female students are treated differently by the evaluators. I develop a new method to answer this question using the peculiarity of the data: for some subjects the scores just below the passing threshold are very rare, apparently because they are often adjusted upwards, so that the student (barely) passes. I check if the probability of such assistance depends on the gender of the student. I also apply the standard approach of checking if any gender does particularly well on oral exams (which are taken locally) compared to their scores on written exams (which are anonymized and evaluated by teachers from other schools). The main finding is that gender does matter. I find that poorly performing female students have a slightly higher chance of passing than their matched male colleagues, although for better students the bias appears to be reversed.