-
Drag and drop the blue circle over the red circle to illustrate the differences in R2 or the difference between correlated and uncorrelated predictors in multiple regression (requires Flash plug-in) .
-
The rectangle in figure 1 (or yellow circle in figure 2) represents a token criterion variable. The area of the rectangle minus the area taken up by the predictors (circles) represents the proportion of error variance (the amount of area that is colored yellowish). The area taken up by the circle(s) represents the proportion of variance that goes accounted for knowing what the predictors are.
-
Note that as the circles overlap more, there is more yellowish area relative to when the circles do NOT overlap as much. The overlap between the two circles indicates the degree to which the two predictors are correlated.