This paper argues that there is a risk of neo-formalism in certain types of empirical research on constitutionalism and judicial review. The risk is driven by the dependence of empirical legal research on quantifiable variables, such as the text of constitutions. The paper outlines the risk through discussion of several examples of neo-formalist empirical research, and proposes some solutions. The paper is relevant to current debates on comparative methodology and the purpose and future direction of comparative constitutional studies.