How to interpret this residuals vs fitted plot for logistic regression using R I am working on a logistic regression on some fundraising data where "gave" is a rare event (approx 3.5%).  My current model has 64% accuracy on test data and an AUC of .604.
When I run the standard plot(model) I am seeing some outliers on each plot that are segmented completely away from the rest of the data set. What is your first thoughts when you see segregation in the plot like this? Is this indicative of a problem or just the nature of the nature of such a rare event?

 A: You did not motivate why proportion classified correctly or residual plots are relevant for the problem you are trying to address.  Assuming your sample size is relevant to the task (loosely speaking, the frequency of "gave" is $\geq 15\times p$ where $p$ is the number of candidate parameters in the model), it may work better to specify a model that is as flexible as the sample size will support (e.g., decide whether to assume linearity for any of the $X$s), and to fit that model.  If you then want to have more comfort that the model is OK (without necessarily being able to do anything about it because the sample size may not allow) you can do a joint test of linearity of all $X$s for which you assumed had linear effects (using e.g. restricted cubic spline functions).  Likewise you can do a global test of all pre-specified interaction effects to assess the additivity of the model.
To describe the usefulness of the model, the $c$-index (AUROC), generalized $R^2$, and $g$-index may be helpful.  Case studies may be found in http://biostat.mc.vanderbilt.edu/wiki/pub/Main/RmS/rms.pdf.
