PaperThe need for more emphasis on prediction: A “nondenominational” model-based approachDavid A. HarvilleAmerican Statistician
PaperAn alternative way to establish the necessity part of the classical result on the statistical independence of quadratic formsDavid A. Harville, Oscar KempthorneLinear Algebra and Its Applications
PaperMaking REML computationally feasible for large data sets: Use of the Gibbs samplerDavid A. HarvilleJournal of Statistical Computation and Simulation
PaperUse of the Gibbs sampler to invert large, possibly sparse, positive definite matricesDavid A. HarvilleLinear Algebra and Its Applications
