Residual plot

A graph which plots residuals against fitted values. It is used to check equal variance assumption of the error terms in linear regression. Residual analysis for logistic regression is more difficult than for general linear regression models because the responses Yi can take only the values 0 and 1. Consequently, the residuals will not be normally distributed. Plots of residuals against fitted values or explanatory variables will be uninformative. Residual plots are generally unhelpful for generalized linear models (where index plots and half-normal plots are preferred).