Читать книгу Elementary Regression Modeling - Roger A. Wojtkiewicz - Страница 12

When there are two independent variables in a model without any interaction term, then the effect of one independent variable does not depend on the value of the other independent variable. An additive regression equation has no interactions, and the effects for any particular independent variable are specified to be the same for any subgroup that one might define.

However, there are instances when the researcher hypothesizes that the effect of one independent variable depends on the value of a second independent variable. An example of an interaction is a hypothesis that says that the positive effect of attending private high school on chances of college attendance is greater for lower income students than for higher income students. The effect of attending private school is hypothesized to be different in the two groups.

Testing for interactions by using regression modeling involves first estimating an equation with additive independent variables and then estimating an equation with an interaction between the independent variables. There are different ways to specify interaction models, and I discuss the various ways in this book. One type of interaction model estimates the effects of an independent variable in different subgroups. A second type of interaction model estimates the differences between subgroups in the effects of an independent variable.