This paper explores the relationship between unionism and quits. Three channels of influence are investigated: unions-collective voice-quits; unions-training-quits; unions-job dissatisfaction-quits. Estimates of each model, using data from the Australian Longitudinal Survey, indicate that unions reduce the probability of quitting via the training effect by 0.5 percentage points, they reduce the probability of quitting via the collective voice effect by 4 percentage points and they increase the probability of quitting via the job dissatisfaction effect by 1.2 percentage points. The net effect of unions is, therefore, to reduce the probability of quitting by around 3 percentage points.