This paper re-examines the returns to computer use using a new matched workplace-employee data from Canada. We control for potential selection using instrumental variables. Results suggest that it is not merely the employee having a computer on his desk, but rather having complementary computer...