Performance evaluations are far from an exact science. Various studies have suggested that the evaluations may be imperfectly designed or that raters have limitations that lead to errors. Perhaps, though, the raters have other goals besides an accurate assessment of their subjects.

That's the thrust behind research by Xiaoye May Wang, Kin Fai Ellick Wong and Jessica Y.Y. Kwong, who look at performance evaluations in peer and non-peer settings. They show that not only do raters' goals affect the outcome of evaluations, but they are applied differently depending on the performance level of the people being evaluated - the ratees. The raters inflate the grades of one and deflate the other according to their goals.

The authors focus on four goals commonly related to performance: harmony, fairness, motivation and identification. The latter is used as the control because it is a standard measure of a person's strength and weaknesses.

In one study, undergraduate students working in groups on a project were asked to submit peer evaluations. The evaluation sheets were manipulated to cover all four goals so as to show how evaluations are adjusted when certain goals are at play.

Under the harmony goal, which aims to reduce conflict, the students significantly inflated the scores of low and medium performers compared to the identification/control goal, but they only minimally inflated the scores for high performers.

Similarly, scores for low and medium performers were inflated under the fairness and motivating goals. However, while the authors predicted the students would deflate the scores for high performers to promote equity or equality under the fairness goal, and to keep the high performers thinking they can do better under the motivating goal, this didn't happen. They inflated these scores, too, but by a much smaller degree.

"We suspect that when it comes to fairness, the raters might have used double standards and increased ratings for low performers according to the equality norm, and not distorted the ratings for high performers according to the equity norm."

"Under the motivating goals, they may have assumed high performers were already highly motivated and so didn't need to receive negative feedback to be motivated."

In a second study, students watched a clip of The Apprentice and evaluated six of the show's team members. This non-peer evaluation gave a different perspective on rater goals, although as with the first study, raters distorted their ratings to achieve certain goals.

The biggest effects were under the fairness and motivating goals. With fairness, the ratings were deflated for high performers but not distorted for low and medium performers. With the motivating goals, scores were inflated for low performers but not significantly changed for medium and high performers. No significant effect was found under the harmony condition, which was unsurprising since they were not rating their peers.

"The ratings were distorted differently for high and low performers in both studies, but the pattern was the same: rating discriminability was reduced," the authors say.

In terms of the practical application of the research, the authors point out that performance evaluations are often used to achieve other goals in management such as maintaining group harmony and promoting group productivity. "Therefore raters have to consider possible detrimental outcomes if accurate performance rating results are released. When rating accuracy isn't the primary goal of raters, they may intentionally distort their ratings to fulfil other goals," they say.