Does Feedback Really Improve Performance?

It frustrates me when things are presented as truths when they are at best opinions, and at worst self-serving fabrications. This is the reaction I had to a recent article in the Washington Post interviewing the author of a Harvard Business Review article about his soon-to-be released book about performance feedback. My goal here is not to attack this author. My goal is to point out things the author is doing that I believe are both wrong and inappropriate. I will do this by critically examining a few quotes the author makes during the interview. The quotes are placed in two categories of things people do when trying to generate market publicity using misleading pseudo-scientific claims.

First, make stuff up

One common way to generate publicity is to make bold statements that are counter to prevailing beliefs. Never mind that many common beliefs are common because they are true. The author demonstrates this technique by leading with the following statement:

We think the thing we should be doing is continuously giving each other feedback. But there’s no research at all that says that leads to greater performance.

This claim is flat out false. There are hundreds of studies looking at the impact of feedback on performance. The overwhelming consensus is feedback improves performance if it is delivered the right way in the right conditions. Note that the part about how feedback is delivered is really important, as feedback delivered the wrong way can hurt performance. Nevertheless, on average feedback does tend to positively impact performance. Here is small sample of findings from this massive body of peer review, non-commercially oriented, empirical research:

“[Based on analysis of 21 studies representing 7,707 employees] nearly all the effect sizes for direct report, peer, and supervisor feedback were positive.” (Smither, London, & Reilly, 2005)
“A meta-analysis (607 effect sizes; 23,663 observations) suggests that feedback interventions improved performance on average” (Kluger & DeNisi, 1996)
“Combining feedback and goal setting is superior in affecting performance to providing goal setting alone.” (Neubert, 1998)

The author then attacks the belief that high performers often share common traits. He uses this to claim that feedback does not work since supposedly there are no common elements associated with effective performance:

“But you actually look at excellent sales people, excellent nurses, excellent doctors, excellent leaders, and the first thing that strikes you is they all look and behave differently.”

This claim reflects something called “restriction of range.” If you compare a sample of excellent performers against one another, you may observe that they do not all act identically. But if you compare a sample of excellent performers with a sample of average performers you will see things excellent performers have in common that make them different from average performers. For example, a study of anesthesia teams found that high performing nurses share common mental models not found in lower performing teams (Burtscher et al, 2011).

The author’s claim that excellent people “all look and behave differently” is like saying elite NBA basketball players all look and behave differently. That may be true when they are compared to each other. But they share a lot in common compared to average high school players. Next to average players, most NBA players are highly similar in the sense that they are much taller and faster, and have far greater mastery of basic skills like dribbling, shooting and rebounding.

Now, discover stuff we already know

Another marketing ploy used to generate publicity is to repackage well-established scientific findings as though they were new research discoveries. The author does this when he starts talking about what companies should do instead of feedback: “There’s only three sources of input that are valuable to a team member: facts, steps, reaction.”

The author goes on to explain that to increase performance you should observe someone’s performance, provide them with facts that will help them be successful, tell them if there are specific steps they should start or stop doing, and let them know how their actions might affect or be perceived by others. This sounds like another psychological process often covered in introduction to psychology textbooks. It is called “providing performance feedback.” The author is just restating several well-established rules for providing effective feedback. Feedback should be specific, describe observable behaviors, and be delivered in a non-judgmental fashion.

I have no problem with people restating the importance of applying well-established psychological principles like the use of behavioral-based feedback. What I object to is making unfounded criticisms about the value of feedback to generate publicity. And claiming to have discovered something new about feedback that is already extensively documented in existing studies. It is damaging to our field as HR professionals, misleading to the business customers we serve, and disrespectful to the researchers who have gone before us,