The post begins with a dissection of a poorly-done article on the subject in the New York Times. This dissection is both analytical and fun to read.
But the most important part of the post is its summary of what psychological research tells us about extrinsic rewards
The problem isn't that all "extrinsic" rewards can backfire, just poorly designed extrinsic reward systems....The basic rule is to never use a stronger motivational system than you need to get the job done.
If praise and grades will do the job then there's no need to implement a token reward system, such as cash or rewards for grades. If a token reward system is needed to motivate a child, then use the least invasive rewards that'll motivate the child to engage in the desired behavior (i.e., learning) and fade out the system as soon as possible. For example, trying rewarding with free time or other reinforcing activities before offering cash, candy, treats, or other tangible rewards.
No prize for the New York Times.