The Role of Theory and Anomaly in Social-Scientific Inference

Designing Social Inquiry, by King, Keohane, and Verba (KKV), deserves praise for many reasons. It attempts, seriously and without condescension, to bridge the gap between qualitative and quantitative political science. It reminds a new generation of students, in both traditions, of some main characteristics of good theory (testability, operationalizability, and "leverage," or deductive fertility). It clarifies, even for the profoundly mathematically challenged, some of the central strictures of statistical inference (why one cannot have more variables than cases or select on the dependent variable, or why it biases results if measurement of the independent variable is faulty). It abounds with practical wisdom on research design, case selection, and complementary methodologies. Perhaps most important, it opens a dialogue between previously isolated practitioners of these two forms of analysis and provokes worthwhile discussion. For all of these reasons and more, the book should be, will be, and-indeed, even in its samizdat formsalready has been widely assigned and read. It is, quite simply, the best work of its kind now available; indeed, it is very likely the best yet to have appeared.' At the same time, I think, Designing Social Inquiry falters in its aim of evangelizing qualitative social scientists; and it does so, paradoxically, because it attends insufficiently to the importance of problemation and deductive theorizing in the scientific enterprise. As natural scientists have long understood (see Hempel, 1966), inference proceeds most efficiently by three complementary routes: (1) making clear the essential model, or process, that one hypothesizes to be at work; (2) teasing out the deductive implications of that model, focusing particularly on the implications that seem a priori least plausible; and (3) rigorously testing those least plausible implications against empirical reality.2 The Nobel physicist and polymath Richard Feynman may have put it best:3