In this paper, we test the hypothesis that computer use will lead to productivity gains only if the firm uses an appropriate set of organizational practices. Detailed data on organizational practices and workers’ compensation are obtained through a Canadian longitudinal linked employer-employee database called the Workplace and Employee Survey (WES). Linked data allow us to take into account both worker and firm unobserved heterogeneity through the estimation of a linear mixed model of wage determination. Our results suggest a small but positive computer-wage premium whose size is related to a set of organizational practices.