Measuring the existence and patterning of white flight (WF) using aggregate data has a long history in the social sciences. In this article, we assess past measurement approaches and identify several technical and conceptual limitations. To address these shortcomings, we propose a new multicomponent approach to detecting WF that requires tracts to exceed a minimum level of loss across four distinct dimensions of population change. Using both simulated and real data, we show that our approach provides stronger protection against common classification errors and behaves more consistently across a broad range of neighborhood types. We conclude with a substantive application that demonstrates the utility of our multicomponent approach to researchers who are interested in assessing the "buffering hypothesis." We find that our multicomponent approach detects higher levels of WF than recognized in previous studies, raising important questions about the formation and long-term stability of racially diverse neighborhoods in the United States.