Implementing police in schools is a common strategy for ensuring school safety, but it is unknown whether, to what extent, and for whom the presence of police in schools affects student arrest rates. Utilizing nationwide data from the 2013-2014 Civil Rights Data Collection (N = 92,620), this study examines how police presence is related to student arrest rates, and whether this association varies by student race/ethnicity and gender. Path models and propensity score matching models show that the association between police presence and arrest rates was stronger for all the groups examined in schools with police, particularly for Black students and boys. This provides support for criminalization theories suggesting that police presence results in more arrests.