Thomas DiCiccio's research broadly involves higher-order methods for accurate statistical inference, as achieved through asymptotic, computer-intensive, and Bayesian methods; areas of application involve linear models, skew distributions, and time series.