Although Pearson correlations are central statistical analyses, correlations based on ranking values require fewer assumptions and can capture patterns that are not linear. Bivariate correlations in SPSS include Spearman's Rho and Kendall's Tau-b. Tau-b was introduced as one of the ordinal measures of association in the Crosstabulation discussion (Chapter 10, Crosstabulation Patterns for Categorical Data). Spearman's Rho works well with values that can be ranked, such as the data for the 201 nations used in this chapter.
The following table contains these two sets of coefficients for the two life expectancy variables and the Percent_Urban field:
The Tau-b coefficients differ only slightly from the Pearson correlations shown earlier, while the Spearman coefficients are noticeably higher. This suggests that while GNI_per_Capita and life expectancy...
