The first part of this report discusses the overall statistical planning, coordination and design for several tar sand wastewater treatment projects contracted by the Laramie Energy Technology Center (LETC) of the Department of Energy. A general discussion of the benefits of consistent statistical design and analysis for data-oriented projects is included, with recommendations for implementation. A detailed outline of the principles of general linear models design is followed by an introduction to recent developments in general linear models by ranks (GLMR) analysis and a comparison to standard analysis using Gaussian or normal theory (GLMN). A listing of routines contained in the VPI Nonparametric Statistics Package (NPSP), installed on the Cyber computer system at the University of Wyoming is included. Part 2 describes in detail the design and analysis for treatments by Gas Flotation, Foam Separation, Coagulation, and Ozonation, with comparisons among the first three methods. Rank methods are used for most analyses, and several detailed examples are included. For optimization studies, the powerful tools of response surface analysis (RSA) are employed, and several sections contain discussion on the benefits of RSA. All four treatment methods proved to be effective for removal of TOC and suspended solids from the wastewater. Because the processes and equipment designs were new, optimum removals were not achieved by these initial studies and reasons for that are discussed. Pollutant levels were nevertheless reduced to levels appropriate for recycling within the process, and for such reuses as steam generation, according to the DOE/LETC project officer. 12 refs., 8 figs., 21 tabs.