ABSTRACT
In this chapter, we continue with our examination and application of variance in the analysis of data, the consideration of the logic of measurement, and the systematic testing of hypotheses in bivariate models. We move on to the foundations of hypothesis testing in cross-national analysis with parametric statistics. As we noted in Chapter 8, the use of these statistics requires that we meet (within reason) specific assumptions about the distribution of scores in a sample before we can appropriately and effectively proceed with our hypothesis tests. Students of cross-national analysis are not always afforded the luxury of having data that are truly interval/ratio, nor are they able to work with sizable samples that minimize the variance differences between ordinal- and interval/ratio-level data. However, through careful combination of both parametric and nonparametric statistics (nonparametric statistics relax assumptions of normal distribution), the student of cross-national analysis can advance our understanding of institutional authority and political behavior through the effective application of hypothesis testing to cross-national data. In this chapter we explore one technique that has, like crosstabs, been a statistical staple for decades and offers important analytical and 245descriptive firepower for the student of comparative politics: the difference of means test. To understand and appreciate this analytical tool, we consider one of the central concepts of cross-national analysis: political culture.
