Manuscript Type Empirical Research Question/Issue While corporate governance research is the beneficiary of advances in research methodologies and statistical techniques, less attention has been placed on variable measurement. This paper draws into question the conceptualization and measurement of CEO duality by highlighting its largely unrecognized instability and the challenges instability imposes on measuring dichotomous variables. CEO duality is widely used in corporate governance research and frequently operationalized dichotomously as a dummy variable. We present examples of the frequent changes in duality within organizations which challenge our current view of CEO duality. Research Findings/Insights We find that the instability of CEO duality in practice varies considerably at both the national and within‐firm levels. We find that a mismatch exists between the current conceptualization of CEO duality, actual patterns of data, and the measures used by governance researchers. The paper draws attention to the limits of conceptualizing and measuring what is seemingly dichotomous data, reviews these in research and in practice, and provides examples, recommendations and assessments of alternate ways existing data can be used. Theoretical/Academic Implications Our results draw into question the reliance on a simple dichotomous conceptualization and operationalization of CEO duality in governance research. Data limitations of corporate governance research may be alleviated by directly assessing stability of duality within firms and reimagining concepts in ways that can be measured using existing data. Practitioner/Policy Implications CEO duality, a legal but discouraged governance structure, may be changed intentionally or result from a variety of temporary firm‐level factors. Assessing the longitudinal patterns in duality and underlying causes for temporary changes in duality should be incorporated into evaluations of firm governance structures.