Moderators are believed to play a crucial role in ensuring the quality of discussion in online political debate forums. The line between moderation and illegitimate censorship, where certain views or individuals are unfairly suppressed, however, is often difficult to define. To better understand the relationship between moderation and censorship, we investigate whether users' perception of moderator bias is supported by how moderators act, using the Big Issues Debate (BID) group on Ravelry as our platform of study. We present our method for measuring bias while taking into account the posting behavior of a user, then apply our method to investigate whether moderators make decisions biased against viewpoints that they may have the incentive to suppress. We find evidence to suggest that while moderators may make decisions biased against individuals with unpopular viewpoints, the effect of this bias is small and often overblown by the users experiencing bias.We argue that the perception of bias by itself is an issue in online political discussions and suggest technological interventions to counteract the discrepancy between perceived and actual censorship in moderation.