Measures of association for ordinal data

Abstract

This chapter concentrates on measures of association for variables measured at the ordinal level. These are PRE measures of association that are similar to lambda. With ordinal data, though, we have more information than with nominal data; in particular we know the ranking of cases. Ordinal measures of association take advantage of this extra information. With lambda, we tried to predict the value an individual case takes for the dependent variable. Since ordinal measurement allows us to rank any two cases ordinal measures of association try to predict the order of pairs of cases.