Scholars engaging in comparative historical analysis rely on observational data. The use of these data to analyze causal factors is affected by a number of problems that follow from studying historical developments. First, scholars often face a dearth of data which inhibits both large-n statistical analysis and in-depth qualitative analysis of individual cases. Second, scholars often confront the problem of historical diffusion, a problem that is in itself aggravated by the absence of good historical data. Against this backdrop, this paper devises a "middle-range" framework based on congruence analysis. This framework includes considerations about case selection in the face of diffusion and about the kind of within-case evidence that allows scholars to analyze the causes of political change. The framework devises a systematic way of converting narrative historical descriptions into within-case binary observations. This conversion allows other scholars to replicate the analysis in a transparent way and to probe the validity of the scoring based on the sources referred to. The framework is illustrated via empirical analysis of the origins and development of representative institutions in the medieval Crown of Aragon.