Using linked employer-employee data from the British 1998 Workplace Employee Relations Survey, we find a positive correlation between workplace union recognition and private-sector employer-provided training. We explore the avenues through which union recognition might affect training by interacting recognition with the closed shop, the level at which pay bargaining takes place, and multiunionism. For non-manual-labor men and women, only union recognition matters. The various types of collective-bargaining institutions have no separate effect. However, the male manual training probability is significantly increased by union presence only through multiple unionism with joint negotiation. In contrast, for women manual workers, union recognition at the workplace has no effect on the training probability.