Community policing is often seen as a way to repair fractured relationships between law enforcement and the public. Despite its relationship-building promise and widespread department-level adoption, individual officers show varying levels of support for community policing which can harm policy implementation. Why are some officers more supportive of community policing than others?Prior research suggests that demographic factors such as the officer's gender, race, age, and education can explain this variance. Across these studies, however, there are several contradicting or non-replicating findings. Conflicting findings may result from differences by department or differences in methodology or temporal variancebut most policing studies focus on a single department. We begin to adjudicate between the possible explanations for conflicting findings using roll-call survey data from 741 officers across three neighboring police departments in 2016 and then a replication with 452 officers from one of the original departments in 2019. We find that experience with community policing consistently influences support for the practice while officer gender, age, and education consistently do not. Other results do not replicate across department or time, although we do find non-replicating significant factors associated with officer support for community policing. Our findings also suggest that departmental and temporal aspects help to explain why policing studies often do not replicate or generalize to other places or contexts.