ABSTRACT
A linear relationship is simple to specify mathematically and simple to represent graphically. This chapter focuses on methods for describing and drawing inferences about a linear relationship between two variables. Many studies involve variables that we would expect to have, at least approximately, a straight-line relationship. In fact, even variables that have a curvilinear relationship can often be transformed into variables that have a straight-line relationship. Straight-line models are one type of a general category of model called linear models. The chapter considers one form of nonparametric correlation: easy-r. Any arrangement of points that puts all the points into two diagonally opposite quadrants will give a easy-r of plus or minus 1.00. This will be true of any monotonic relationship (one in which the direction of the relationship stays the same throughout the range of scores), of which linearity is simply one example. The chapter also considers two more powerful nonparametric measures of correlation.
