A **residual plot **is a graph used to demonstrate how the observed value differ from the point of best fit. A residual plot will have the appearance of a scatter plot, with the residuals on the y-axis and the independent variable on the x-axis.

A residual plot is used to determine if residuals are equal, which is a condition for regression. To know if the residuals are equal, you will be looking for a residual plot that has a random homoscedastic distribution, such as graph (a) in the figure to the right. You do not want a residual plot to have a pattern or that is heteroscedastic, such as graphs (b) to (f). To learn more about the six plots found in this figure, scroll to the section titled ‘Predicted Values and Residuals’ (about halfway down) on this page.

A simple definition, some other examples, and a small quiz can be found here. For the more visual learners, this is a handy video. Starting at 2:12, the author discusses many different types of residual plots and addresses why they may look the way they do, and why or why not they are useful.