James Stark
Decision Tree Exercise

One of the most useful exercises I do in my negotiation class works like this:

1. Students receive an out-of-class negotiation simulation (involving an alleged breach of contract) packet containing legal information as well as a confidential client profile.
2. They spend about a week with the problem and then conduct videotaped negotiations on their own.
3. They come back in seminar and discuss the negotiation--processes, strategies, outcomes. Students learn the confidential facts in the opposing party's profile. By design, I spend comparatively little time in this class talking about the law pertaining to the problem. (Most students say, "there's law on both sides" and "it's not productive to discuss the law in a negotiation" anyway.)
4. For the next week's assignment, I give them reading materials on "decision tree" analysis. I construct a decision tree for the case, asking them to assume out-of-pocket damages of $50,000, and ask them as a homework assignment to make a prediction of their chances of prevailing at trial based on the facts and the legal elements of the case (contract? breach? forseeable damages?). From this analysis, they are asked to come up with an "expected value" for the case.
5. When we reconvene the following week, I immediately divide the students into two teams: plaintiff's lawyers and defendant's lawyers, and ask them to make their predictions together as a group, coming up with consensus percentage figures for each branch of the tree, and a consensus expected value.

I've done this two years in a row, and each year the plaintiff lawyers' settlement value has been more than twice the defendants'! The following week, I have them read stuff on egocentric and other cognitive biases in negotiation.

Hope this is a useful approach for the context in which you work. I suspect that to demonstrate the full effects of partisan bias, students have to live with a case (real or simulated) for a while.