Numerous studies have documented a positive association between information technology investments and business- and establishment-level productivity. Most of these studies, however, rely on empirical specifications that over-represent small businesses. In this paper, we revisit one piece of evidence, the Computer Network Use Supplement to the 1999 U.S. Annual Survey of Manufactures, which has previously been used to show that there is a positive relationship between computer networks and productivity in manufacturing plants. We show that this is only true for small- and medium-sized plants, and that for larger plants the relationship is negative. We give critical consideration to alternative methods for weighting these data, and show that employment-weighted estimates indicate the presence of a computer network has, on average, a negative relationship with the productivity of employees.