Residual plots are traditionally used to assess the fit of a regression model, yet can be difficult to interpret when the response variable is binary. This difficulty becomes compounded when covariates have a hierarchical or multilevel structure. An alternative graphical procedure is proposed for visualizing goodness of fit in such settings. The methodology is illustrated with an analysis of individual-level and county-level effects on sentencing practices across the U.S.
Back to Research page.
Send me e-mail at ipardoe at lcbmail.uoregon.edu
Last updated: August 27, 2002
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.
© 2002, Iain Pardoe, Lundquist College of Business, University of Oregon