This study examines the relationship between union membership and (i) the incidence of training, (ii) the degree to which training is transferable to firms other than the one providing the training and (iii) the degree to which workers perceive that training improves job performance. Using data from the Australian Bureau of Statistics, I find that union members are more likely to receive employer‐sponsored training than their non‐union counterparts. I also find that male union members are more likely than non‐members to report that training improved job performance. Union membership was not related to transferability of skills between employers.