Student Evaluations of Instruction (SEI) are an important issue in countries like the USA, where the evaluation results can impact professional promotion chances and salary of faculty. According to Seldin [11], the percentage of American colleges using SEI grew from 29% in 1973 to 68% in 1983 and to 86% in 1993. Consequently, the adequacy of SEI has been examined extensively, and many statistical studies have been published. Non-instructional factors, which cannot be influenced by instructors, may bias the evaluation rating and should be identified and eliminated for a fair comparison. But in many cases, a mere linear regression of SEI on such potential factors is not adequate.