ABSTRACT
Estimating the variance. Ordinary Poisson regression estimates. Rescaling the variance using the Pearson dispersion statistic. Quasi-maximum likelihood. The robust variance, also called the Huber, White, survey, or sandwich estimate. Generalized estimating equations. Bootstrap methods, including the normal, percentile, bias-corrected percentile, bias-corrected and accelerated, and bootstrap-t methods. Permutation and randomization. Comparisons using real data.
