[hedgew]

Like so many studies about bias and heuristics, this one probably does not generalize well beyond trivial tasks. I trust that the research was executed well, but I doubt the results say anything definite, other than that in trivial tasks performance can't be increased by increasing rewards.

Examples of the "work tasks" in the studies are: adding numbers and tapping a key as fast as possible. You can probably imagine quite well that it's very easy to hit your maximum performance in these tasks and you can't do much to exceed it no matter what the reward. It seems a bit simplistic to use these studies to claim that bonuses can't increase quality.

The article also completely disregards that bonuses can be used to improve principal-agent alignment, and probably many other factors that are apparent in the more complex, real world that exists beyond tapping a key as fast as possible.