Original paper here: http://www.umass.edu/preferen/You%20Must%20Read%20This/herrm...

The explanation in the linked article isn't fully convincing, IMO. The paper speculates about several more explanations at the end, and has some more data to support them, e.g. people seemed to be punishing low-contributors in an effort to make them contribute more.

If only the experimenters had conducted some interviews...