Thursday, November 1, 2001

Letter to the Editor
Teacher evaluations not correct use of statistics

Two commentaries in the Oct. 26 Skiff questioning the usefulness of the current professor evaluation system caught my attention. Some students criticized them, and some supported the evaluations.

Regardless of the different opinions, what I found shocking was the lack of mention to the fact that the current evaluation forms used by the university are statistically incorrect, and the data being used to evaluate professors is simply illogical.

The current evaluations used at TCU are a type of statistical analysis that asks students to rate on a scale of 1 to 5 the qualitative attributes of a class or a professor. Using correct statistical analysis, the answers would be tabulated, and then the percentage of answers would be evaluated.

However, under the current method, instead of measuring the percentage, they tabulate the results and use an average instead. Using averages for qualitative analysis is simply bad use of statistics.

In reality, which piece of information makes more sense: the fact that a professor got an average of 3.5 in their class preparation, or the fact that 70 percent of the students thought the professor was well prepared for class?

So, until the university fixes the current way evaluation results are inappropriately used, further changes in the forms are useless. Maybe TCU students and administrators could benefit a little from taking DESC 20153, and then design new forms that reflect not only student input and opinions but also correct use of statistical analysis tools.

— Raquel Torres Carvajal
senior e-business major