From+Case+Studies+to+Social+Science++A+Strategy+for+Political+Research

Overall point: Bates is arguing that description (as in qualitative research), explanation, and confirmation (which tends to require quantitative study) are complementary, not competing, that all are involved in comparative politics, and that all are consistent with rational-choice assumptions. More detailed summary: There are three levels of understanding--”apprehension or //verstehen.”// This is knowing a situation richly, having been there or as if having been there, and gaining insight. This is the realm of ethnographers and historians. The emphasis is on description and interpretation, but not explanation. The next level is “explanation.” This involves the “therefore” component--being able to understand the logical links--how one thing leads to something else. It requires abstracting away from the situation to some degree and looking at multiple cases in order to uncover patterns. Game theory falls into this category. Despite the methodological debates, these two forms of understanding are complementary. Apprehension///verstehen// helps provide intuitions that spark explanatory hypotheses and serve as a “bullshit meter” (author’s phrase). Interesting quote on the complementarity of rational choice and qualitative work: “The assumption of rationality provides a source of empathy that enables the analyst to occupy the position of those whose behavior she hopes to understand.” The third level of understanding is confirmation. You need “large-N” statistical analysis to convince oneself and others that the theoretical “therefore” linkages are correct. Here, Bates switches to an account of his own research to illustrate the point. He studied coffee growers in Uganda and Kenya, asking why they did not produce as much coffee as they were able to and maximize profit as economic theory would predict. Existing explanations had focused on a “culture of affection” and not wanting to outdo their neighbors. Bates doubted this theory, and used the two cases and changes over time (before and after major regime changes in both countries) to demonstrate that the farmers were making choices in response to changing political, economic, and military conditions, rather than acting according to static features of “culture”. So a “small-N” comparison helped with explanation, but it was not enough to create “conviction.” He went to Colombia and found that peasant coffee growers were treated much better--a result, he thought, of coffee growers being a substantial swing vote in a political system of two-party competition. So he looked at the regimes in Colombia over time (that is, increasing the number of cases to compare, but all within Colombia), hypothesizing that when there was a two-party system, growers fared quantifiably better than when there was a single party system. His results confirmed the hypothesis. He next went to a larger s a larger set of cases to test his hypothesis that the existence of two-party or multi-party competition correlated with a more free-market economy, with less bureaucratic control of markets. By using a large set of cases in Africa (46 countries over a 26-year period), he was able to control for other factors that might account for this difference. This illustrates how “large-N” statistical methods help confirm the insights gained from immersion in single cases and the theories and hypotheses that emerge from “small-N” comparisons.
 * Bates--summarized by Suzanne **