Research Summary This research addresses firms' use of external knowledge sources to develop patented inventions and explores the validity of patent citations as an indicator of interfirm knowledge flows. By comparing patent citations with primary data reported by the inventors, we uncover systematic measurement errors in patent citations and show that they depend on the firms' patent strategies (e.g., to reduce the risk of imitation or litigation), the source of knowledge employed (e.g., competitors, users), the technology of the underlying invention, and the institutional characteristics of the patent system. Our findings about the role of these factors in external knowledge sourcing and citing propensity highlight the importance of firms' strategic behavior and offer novel insights for the use of patent citations as an indicator of knowledge flows. Managerial Summary Firms' open innovation strategies rely on the sourcing of knowledge from other organizations. Tracing these knowledge flows is difficult, such that the empirical research on this matter typically uses citations that patents make to prior art in order to track them. However, patent citations might be added also for reasons other than the actual transfer of knowledge. We use primary information from a large survey of inventors to assess the accuracy of patent citations to measure knowledge flows, and we find evidence of measurement errors that depend on the applicants' patent strategies, the type of knowledge sources used, the filing jurisdiction, and the technology of the underlying invention. We offer insights to evaluate the settings in which patent citations are a reliable measure of knowledge flows.