This article describes a systematic method for reducing the number of questions in a network instrument, without losing key elements of the analytical structure present in the larger questionnaire. Our goal in this process has been to conserve the instrument's theoretical structure and to preserve its analytical power, while reducing unnecessary redundancy. The process results in a revised data collection instrument which creates less of a burden on the networks from which we collect data, and a reduced number of analyses that must be completed on the data set, without sacrificing richness of data or strength of interpretation.