Animated interpolation between a marginal scatterplot
and an added-variable plot

These plots show the relationship between a marginal scatterplot and an added-variable (aka partial regression) plot by means of an animated interpolation between the positions of points in the two plots. It uses simple linear interpolation between the two sets of points, of the form

  xy <- xy1 + α * (xy2-xy1)

where α ranges from 0 to 1.

It shows how the slope of the regression line, the data ellipse, and positions of points change as one goes from the marginal plot (ignoring the other variable) to the AV plot, controlling for the other variable. In each panel, the 3 points farthest from the centroid are labeled.

The data used here are Duncan's data on the relation between occupational prestige and income and education in those occupational categories, fitting the model

  duncmod <- lm(prestige ~ income + education, data=Duncan)