Two forms of a teacher rating questionnaire, Student Reaction to Instruction, were administered to college students. The regular format used category scaling; the 631 responding students selected a number between one and five. Experimental "ratio production (multiply-divide)" evaluations were also completed by 26 subjects along with the regular form. The ratio production form required students to judge whether the particular dimension was more than, exactly the same as, or less than the specified standard (other teachers at this university). If "exactly the same as" was not an appropriate response, the students filled in a number to represent the perceived ratio (two times more effective, four times less effective). The multiply-divide scale produced ratios and eliminated some of the problems associated with traditional category scaling: confusing interpretations and highly skewed distributions of means. The tendency to rate instruction high, rather than low, was apparent with both methods; however, with the multiply-divide scale, there was a greater tendency to rate an instructor "the same as" rather than "average". Results were illustrated graphically, on semi-log scales. Respondents preferred the more commonly used category scale form to this new method; further research was recommended. (GDC)