Evaluations

Evaluations provide a formal assessment of someone's job performance often with recommendations on how to improve and further developmental goals. Feedback is often a part of evaluations, but is distinct. Many organizational behavior experts suggest de-coupling evaluations and feedback as much as possible. [HBR]


Facts and data

  • Unconscious beliefs about what leaders look like can affect evaluations. Studies have found that people underrepresented in business leadership are often deemed less suitable for leadership roles. For example, a 2008 study asked participants to read a story about a male CEO and then to rate the CEO on his effectiveness as a leader. When the CEO was described as white, he was perceived as a more effective leader than when he was described as black. The story was exactly the same [Paradigm]

  • Biases can lead to gender double-standards. One researchers who looked through many performance evaluations found the same behavior was treated differently from men than women. One review read “Heidi seems to shrink when she’s around others, and especially around clients, she needs to be more self-confident.” But a similar problem —­­ confidence in working with clients —­­ was given a positive spin when a man was struggling with it: “Jim needs to develop his natural ability to work with people.”[HBR]

  • Bias creeps in most the evaluation process is ambiguous and open-ended. Performance evaluations without structured intermittent feedback and evaluation sessions tend to rely on the supervisors impressions. Impressions are very susceptible to unconscious bias. Additionally, evaluations that lack clear metrics tied to job outcomes also invite impressions rather than objective data [Paradigm]



Champions

I had no idea of what I was doing, and it was a nightmare. The line of work I was assigned to was really complicated, and I didn’t think I could last here. The conversations I had with my line manager were very honest — “This is where you’re technically capable, but also this is where you’re not” — and I could then look at the cold facts of the ratio between the two. We had a chat about what support would and wouldn’t be available to help with some of those gaps, how much I’d have to learn and how fast I’d have to learn it, and we compared it to some of the things I’d done previously, and it came out it’d probably work out. It’s really worth having someone say, “You did that really well, your technical work there was great, but you didn’t explain it in a manner that your audience could understand.” Very honest feedback helped me to now have the confidence to know what I can do and try to swim when thrown in at the deep end. ~ Jennita, Process Engineer (source)


Hear it FirstHand

There was a woman who I supervised who wasn't performing up to the standards I had. I tried to bring this up to her at one point, but she seemed upset by what I had to say. After that I got nervous about giving her feedback because I didn't want to make her feel bad. But she wasn't performing well either and so I was stuck with this situation where I really needed her to improve but I didn't know how to tell her. It wasn't good, and I realize now that even though I was trying not to hurt her feelings, I also was not helping her career. 


Take Action

  • Use hard data and clear metrics for evaluations. Make key decisions based on hard data and less on subjective, qualitative elements, such as comments on a candidate’s personality or personal circumstances [BCG]

  • Evaluating employees in small groups and comparing them to each other reduces bias. Researchers gathered data on how well 100 experiment participants, "employees", performed at math and verbal problems (polls of gender attitudes show most believe men are better at math and women better at verbal tasks). They then threw out the scores of all participants except those who got mediocre test results. Next, they asked 554 other participants, "employers", to choose from this smaller set of "employees" (including knowing their gender) and choose which they thought would do best on another round of tasks. When "employers" were given only one "employee" at a time, gender stereotypes played out – male candidates were chosen for math t asks and female candidates for chosen for verbal tasks. When "employers" were given both a male and female candidate at the same time, this stereotype disappeared [HBS Working Knowledge]


pjford