Non‐cognitive dispositions have recently become psychological constructs of interest in the education, economics, philosophy, sociology, and psychology literature. In this article, I draw the distinction between property‐first and measurement‐first approaches to understanding the meaning of a particular non‐cognitive disposition theoretical term, ‘grit’, as well as the relationship between this term and its corresponding measurement procedure. I argue that, depending on which approach one takes, one is confronted with different norms for judging the success of any disposition‐measurement procedure pair, meaning that the difference in philosophical approaches may have practical import. I argue that the measurement‐first approach is the one educators and educational policy‐makers ought to adopt towards grit as it is less likely to lead to over‐valuing grit in situations where a gritty disposition may be harmful, better respects scientific virtues and values, and provides a clearer understanding of grit than the property‐first approach.