Book Review of 'Statistical Modeling and Inference for the Social Sciences.

Summary 1st line: For the past twenty years, numerous works have been devoted to "mixed methods" in political science. Despite the diversity of the procedures developed, their promoters are generally motivated by the same observation: qualitative approaches, applied "in depth" to a small number of trajectories, could usefully serve to complete, qualify, or even challenge theories demonstrated by quantitative methods on larger populations of "cases. Thus, although they allow for the simultaneous examination of many units, they would overlook factors that are potentially decisive for the situations under consideration, but that do not lend themselves well to modeling - such as institutional inertia, the role of representations or the legitimacy conferred on certain groups.