ABSTRACT
The main focus in this chapter will be upon the utility of regression as a tool to analyze error (that is, residuals) in our models, and from the model error, to elaborete and refine our cross-national analysis through residual analysis. As we have seen in several of the previous chapters, the use of residuals—or the examination of outliers relative to the sample distribution—is an effective way to isolate unique cases within a cross-national sample and from this to refine our theory and sharpen our conclusions. Residual analysis is more than a mundane and tedious technical exercise within quantitative analysis. Indeed, it lies at the very core of comparative political inquiry s contribution to general theory building. Residuals are the signposts guiding us to new avenues of discovery and theory refinement. They offer the possibility of conceptualizing our political world in ways we had not thought of before, because they are the empirical representation of the unexpected—the inexplicable. “Theory building,” notes Lawrence Mayer, “consists of specification and analysis of factors that were part of the unexplained residuals in existing explanatory theory.” 1 Thus, in regression analysis, we have a powerful tool not merely for hypothesis testing, but for teasing from cross-national data further insights into patterns among countries which illuminate pathways for future research.
