ABSTRACT
If we are simply interested in the nature of association between variables (positive, negative, or none), then correlation analysis should be used (Chapter 8). Conversely, if we hope to show that one variable or group of variables explains variation in another variable, or to simply model this relationship, we would use regression analysis. Regression analyses require the designation of one response (Y) variable and one or more quantitative explanatory (X) variables (i.e., predictors).* For instance, to model the heritability of plant height, we could measure seedling height at maturity (Y), and mean parental height (X), and then “regress” Y on X.
