ABSTRACT
This chapter examines more depth about multiple regressions and explains some concepts a little more fully. It explores concepts several different ways to ensure that at least one explanation makes sense to every reader. Among other things, residuals are useful for diagnosing problems in regression, such as the existence of outliers, or extreme values. The residuals are what is left over or unexplained by the regression equation. The regression line, with simple regression, is simply a line connecting the predicted Y’s for each value of X. With multiple regressions, however, there are multiple regression lines. But wait: if the regression line is equivalent to the predicted scores, then the predicted scores are equivalent to the regression line. Students will sometimes hear simple or multiple regressions referred to as least squares regression or ordinary least squares regression. The reason is that the regression weights the independent variables so as to minimize the squared residuals, thus least squares.
