ABSTRACT
This chapter introduces multiple regressions, with two independent variables and one dependent variable. It examines a regression designed to determine the effect of time spent on homework on grade-point average, controlling for parents’ level of education. The regression equation was statistically significant. Unlike simple regression, with multiple regressions it is possible for the overall regression to be statistically significant but to have some independent variables be nonsignificant. The chapter discusses the pros and cons of interpreting standardized versus unstandardized regression coefficients. The path diagram, or path model, shows the variables in the multiple regressions in rectangles. Arrows, or paths, are used to signify regression coefficients, and the curved, double-headed arrow between the two predictor variables represents the correlation between them. This model shows the standardized coefficients; it would also be possible to use the unstandardized regression coefficients. The discussion makes obvious the chief advantage of multiple, over simple, regression: it allows controlling for other relevant variables.
