As networks of all forms become ubiquitous, the network-based information they generate is increasingly being used in a wide variety of analysis tasks. In organizations, social network analysis techniques are being applied to a number of domains, particularly the understanding of knowledge stocks and flows. Because this information is generated from large data sets, computerized visualizations of it are very helpful for accomplishing these complex tasks. This paper presents a model for evaluating the effectiveness of network visualizations based on theories of cognitive fit, working memory capacity, and information load. The model was empirically tested in two experiments using two types of data visualizations from two different social networks. Results support the theoretical model, illustrating that variations in cognitive fit and working memory interact. Findings suggest that visualizations can enable superior outcomes when they are designed to support this interaction.