| Psychologists have continually researched the effects of rating scale format manipulations on traditional rating errors (e.g., halo, leniency). Rather than merely addressing these classic rating errors, the present study attempted to determine whether format differences affect the amount of sexual bias exhibited by performance ratings. The subjects' task involved evaluating the performance of a hypothetical professor. One hundred twenty subjects observed a male or female performing the job of professor of home economics or civil engineering. Subsequently, subjects evaluated performance via a graphic scale, Behavioral Expectation Scale, or Behavioral Observation Scale. Results indicated that scale format has an effect on the amount of sexual bias exhibited by performance ratings. Behavioral scales (BES and BOS) did not exhibit bias, while the graphic format was characterized by high levels. Results are discussed in terms of performance appraisal implications. |
