Research Design: Toward a Realistic Role for Causal Analysis

摘要: For a half-century, sociology and allied social sciences have worked with model of research design founded on distinction between internal validity, the capacity designs to support statements about cause effect, external extent which results from specific studies can be generalized beyond batch data they are founded. The is conceptually useful has great pedagogic value, that is, association experimental random sampling validity. advent potential outcomes causation, by emphasizing definition causal effect at unit level heterogeneity effects, made it clear how indistinct (and interpenetrated) these “twin pillars” design. This theme this chapter, inveighs against idea hierarchy desiderata, inference peak. Rather, I adopt typology Leslie Kish (1987), advocates an appropriate balance randomization, representation, realism, illustrate all three elements not just validity mechanism) integrated aspects meaningful analysis. What analysis? It depends first foremost getting straight why we doing what doing. Understanding something happened may tell us lot will happen if were actually do something, but necessarily so.