PARTIAL RESIDUAL PLOTS IN GENERALIZED LINEAR-MODELS

In this article we explore the structure and usefulness of partial res idual plots as tools for visualizing curvature as a function of select ed predictors xa in a generalized linear model (GLM), where the vector of predictors x is partitioned as x(T) = (x(1)(T),x(2)(T)). The GLM e xtension of CERES plots is discussed, but to a lesser extent. The usef ulness of these plots for obtaining a good visual impression of curvat ure may be limited by the specified GLM, the link function, and the st ochastic behavior of the predictors. Partial residual plots seem to wo rk well when modeling is in a region where the conditional mean of the response given x stays well away from its extremes so that the link i s essentially linear, and E(x(1)\x(2)) is linear in x(2). CERES plots, however, require only the first condition. The behavior of these plot s is contrasted with their behavior in additive-error models.