The mechanism through which outsourcing favourably impacts on workplace performance, particularly productivity, is still unclear. I explore the hypothesis that it does so by impacting workers’ training. I use AWIRS-1995, a matched employer–employee survey that reports ample information on the extent of technology and organizational change in Australian workplaces. I find that there is a positive and significant impact of outsourcing on training when I do not control for the correlation between ununobservable factors in these two binary outcomes. However, once I control for this correlation using a bivariate probit estimator, the training impact of outsourcing becomes negative. I then assess the sensitivity of the outsourcing effect to endogeneity by using the method advocated by Altonji et al. (J Polit Econ 113(1):151–184, 2005) to find that this latter result persists even in the presence of a low correlation between unobservables.