rationalskepticism.org

Mr.Samsa wrote:Cool video, the content is bullshit though.

You can find the study here. The problem is that they are assuming that the value of the incentives increases exponentially; that is, they are assuming that the difference between $1 to $2, is the same relative difference as $100 to $200, and $1000 to $2000. However, we know that this is wrong as value increases hyperbolically. The relevance here is that whilst it makes sense to work harder for a "Very Good" performance when your options are $10 and $20, there is less incentive to do so for $100 and $200 - essentially because even if you fail, you get a fair bit of money anyway.

Their claim that "cognitive tasks" somehow violate behavioral laws is also bullshit. What they should have concluded was that the more arduous the task, then the less value incentives have. We already knew this though, we created the laws that determine this behavior over 40 years ago and if we plug the data in from this study, I'm sure we'll find that people behaved exactly as predicted. To highlight this point, in their study their "mechanical task" was to alternate between pressing the 'v' and 'n' buttons on a keyboard, and the "cognitive task" they were given a set of 20 matrixes (with 12 numbers in each) and their task was to find two numbers that would add to 10.

Their general conclusions are kind of right, but they've horribly misunderstood their data, made numerous unfounded assumptions and spectacularly failed at doing a basic literature search before flying all the way to India to "discover" something that was done and dusted over 40 years ago.

Ah economists, always a few decades behind the science...