DataDesk is a powerful tool developed by Paul Velleman for exploratory data analysis. This example is about using color regression to examine how a regression model of two variables is conditional upon a third variable.
In this regression model there are four variables. The dependent variable is posttest scores. The independent variables are pretest scores, cognitive test scores and integration test scores. As shown in the scatterpot, when the cognitive ability is low, the pretest scores and the posttest scores have a negative relationship. The relationship is positive on the condition that the cognitive ability is medium or high. You can view an animated demo or the slide show demo of the procedure.

What would the relationship between the pretest and posttest scores look like at different levels of integration test scores? You can use this dataset and explore it in DataDesk or other statistical software packages.