Here’s a little experiment for your next staff meeting. All you need is an open space about ten feet long and maybe three feet wide, two coins, and a flip chart.
Once you’ve cleared the space, set a target on the floor at one end of the ten-foot length. The target can be a trashcan, a book, a purse … anything to mark a fixed location on the floor.
On the flip chart, write down four categories:
Now have one of your colleagues stand at the end of the ten-foot space farthest away from the target and facing away from it. Give her the two coins. Ask her to take one coin and throw it over her shoulder, trying to get it as close as possible to the target.
Observe where the coin lands. If it’s close to the target, praise your colleague lavishly: “That’s great. You’re obviously a natural at this. Keep up the good work.” If the coin falls far from the target, criticize her equally lavishly: “That was awful. You’re just lame at this. You better buck up.”
Now have your colleague throw the second coin and observe whether it’s closer or farther away from the target than the first coin. Now you have four conditions:
Now repeat the process with many colleagues and watch how the tick marks grow. If you’re like most groups, Category 3 (C+) will have the most marks. Conversely, Category 1 (P+) will have the fewest marks.
So, we’ve just proven that criticism is more effective than praise in improving performance, correct? Well, not really.
You may have noticed that throwing a coin over your shoulder is a fairly random act. If the first coin is close, it’s because of chance, not talent. You praise the talent but it’s really just luck. It’s quite likely – again because of chance – that the second coin will be farther away. It’s called regression toward the mean.
Conversely, if the first coin is far away, it’s because of chance. You criticize the poor effort but it’s really just luck. It’s quite likely that the second coin will be closer. Did performance improve? No – we just regressed toward the mean.
What does all this prove? What you tell your colleagues is not the only variable. A lot of other factors – including pure random chance – can influence their behavior. Don’t assume that your coaching is the most important influence.
However, when we do studies that control for other variables, praise is always shown to be more effective at improving performance than criticism. I’ll write more about this soon. Until then, don’t do anything random.
(Note: I adapted this example from Daniel Kahneman’s book, Thinking Fast and Slow. I think Kahneman may have adapted it from Edwards Deming’s experiment involving a fork and different colored balls.)