The quantitative study of comparative politics is often described as a data-driven enterprise. Employing an original dataset of comparative politics articles published in leading academic journals between 1989 and 2007, this article offers the first empirical analysis of data usage in comparative research. Tracing potential biases induced by data dependence, it assesses the structure of quantitative comparative research (by year, research design, geographic focus, and subject area), the use of country-specific and region-specific datasets, the introduction of original data, and the degree of concentration in data usage. Its empirical findings question cherished assumptions about the structure of the discipline.