Two approaches (Score 1)

Different professors ran their team projects in different ways, which can be grouped into two categories:

1. The students were broken up into assigned groups, usually by something arbitrary like the spelling of their last names. This produced a mixture of competency levels in each team. The outcome was as described in other posts, where the people who cared the most about doing well in the class carried the others. (I especially liked the one post that called a person who did all the work a "doormat" -- it's easy to make someone else into a doormat when you have nothing to lose.) These projects were effectively no different than working alone or with one other person.
2. The students were given the option of doing the project alone or in teams. Based on my past experiences, I initially wanted to work alone, but something really interesting happened: other students in the class approached me and asked me to join their team. I knew them to be some of the better students around, so I agreed, and we spent some very productive time working together on the projects. I got a lot more out of that course than I would have if I had chosen to work alone.

The moral of this story: the battle for successful teams can be won and lost in the initial selection of the team members. If you can't find good colleagues, you are better off working alone. If you can find good colleagues, you're a fool if you don't collaborate with them.