ABSTRACT
This chapter discusses how to measure three types of relationships: those among ordinal and nominal variables; those that involve dichotomous variables; and those that involve several variables simultaneously. A cross-tabulation is to the analysis of ordinal variables what a scattergram is to the analysis of interval variables. A dichotomy is a classification that involves only two categories: for example, gender, referendum vote, or participation. Such classifications have a useful property, which at first glance appears to be a bit mysterious. To handle the problem of describing the relationship between a continuous independent variable and a dichotomous dependent variable, statisticians have developed an alternative model that is found in two very similar variations, called logit analysis and probit analysis. The technique of multivariate regression looks at the pattern in several variables among all observations and to estimate what the relationship between the dependent variable and any particular independent variable vari.
