This authoritative new volume treats a wide class of distributions that constitute plausible alternatives to normality -- such as short- and long-tailed symmetric distributions and moderately skewed distributions -- all having finite mean and variance. Robust Inference illustrates the appropriateness of various robust methods for solving both one-sample and multisample statistical inference problems ... develops Laguerre series expansions for Student's t and variance-ratio F statistic distributions ... analyzes normal and nonnormal distribution efficiencies ... works out modified maximum likelihood (MML) estimators based on type II censored samples for log-normal, logistic, exponential, and Rayleigh distributions ... uses MML estimators in constructing robust hypothesis-testing procedures ... considers the specialized topics of regression, analysis of variance, classification, and sample survey ... discusses goodness-of-fit tests ... describes Q-Q plots in a special appendix ... and much more. An outstanding, time-saving reference for theoreticians and practitioners of statistics, Robust Inference is also an excellent auxiliary text for an undergraduate- or graduate-level course on robustness.