Summary

We can analyze binary data using a binomial GLM. However, we don’t use the ratio of the residual deviance to residual DF to diagnose overdispersion or underdispersion. Instead, we use the binnedplot() function from the arm package to examine the binned residuals. Otherwise, the analysis is interpreted in a similar way to a GLM of binomial count data because the GLM uses the binomial distribution and the logit link function.