A Graphical Diagnostic for Variance Functions

Iain Pardoe and R. Dennis Cook

Abstract

This paper proposes diagnostic plots for regression variance functions. It shows how to extend graphical methodology that uses Bayesian sampling for checking the regression mean function to also check the variance function. Plots can be constructed quickly and easily for any model of interest. These plots help to identify model weaknesses and can suggest ways to make improvements. The proposed methodology is illustrated with two examples: a simple linear regression model to fix ideas, and a more complex study involving count data to demonstrate the potential for wide application.

Keywords: Bayesian methodology; Diagnostic plot; Marginal model plot; Model criticism; Posterior predictive distribution.

Back to Research page.


Send me e-mail at ipardoe at lcbmail.uoregon.edu

Last updated: January 12, 2007


The views and opinions expressed in this page are strictly those of the page author. The contents of this page have not been reviewed or approved by the University of Oregon.

© 2007, Iain Pardoe, Lundquist College of Business, University of Oregon

Valid XHTML 1.0! Valid CSS!