ROBUST ANOVA AND ANALYSIS OF DESIGNED EXPERIMENTS

RESEARCH IS PROPOSED ON ROBUST METHODS FOR ANALYSIS OF VARIANCE (ANOVA) AND ON STATISTICALLY DESIGNED EXPERIMENTS (SDE), WITH A VIEW TOWARD IMPLEMENTATION IN MODERN INTERACTIVE GRAPHICS WORKSTATION ENVIRONMENTS. THE FOLLOWING ROBUSTNESS CONCEPTS AND TOOLS WILL BE USED: INFLUENCE FUNCTIONS, HIGH BREAKDOWN POINT ESTIMATION, MIN-MAX BIAS ROBUST ESTIMATION, AND MEAN-SQUARED ERROR EFFICIENCY ROBUSTNESS. THE PHASE I RESEARCH WILL FOCUS ON FOUR PROBLEM AREAS:(1)ROBUST ESTIMATES OF SCALE; (2)ROBUST ESTIMATES OF SCALE FOR EFFECTS CONTRASTS AND ERROR TERM IN FACTORIAL AND FRACTIONAL FACTORIAL DESIGNS; (3) ROBUST METHODS FOR VARIANCE COMPONENTS MODELS (ONE-WAY LAYOUT IN PHASE I); AND (4) INITIAL DESIGN OF INTERACTIVE GRAPHICS AND USER INTERFACE.