This unit will actually be presenting a review of many of the things we have already discussed, but we will be using some new kinds of examples and in some cases looking at the material from a slightly different point of view. We will also be emphasizing once again the distinction between normative and descriptive and trying to be more explicit about the different ways in which we evaluate normative and descriptive models of decision making.

Below is a check list of the important topics that we will be covering, but the order here is not exactly the same as the one we'll follow in lecture.

Meta-decisions: Suppose we are laying out a matrix for the decision as to whether to go to summer school or to work. Let us call that the primary decision on our agenda. But as we do this, we may face many decisions at the meta level (where meta means above or beyond): Are there other options we haven't included in the matrix, such as maybe looking for a non-salaried internship? Is now the best time to be making the primary decision or should we decide to wait until after exams? (Which doors might open in the meantime - and which might close?) Should we be proactive in collecting more information relevant to the primary decision or would it really be worth it?

When we move beyond the stance of taking the matrix as given, all sorts of interesting new issues arise. To address some of them we can just re-apply rational decision theory at the meta level. So,for example, we will look at how one might use RCT to decide whether it is worth it to collect more information relevant to the primary decision we started with. A key concept here is the cost of information.

Sometimes it's good to step back ( or up) to the meta level but we can't always do this or we'll get into an infinite regress of trying to decide whether to decide whether to decide.... This is one example of what philosophers call the problem of hyper-rationality. (Yes, folks, you can be too rational, at least in a formal sense of the term.)

Decision traps: In Unit 2, we already looked at various ways in which actual decision making behavior differed from the normative account of decision making provided by RCT.Here we'll review some of those, especially the phenomena of anchoring and framing, but use the sorts of examples found in popular books on decision making instead of focusing on the research findings. We'll talk in detail about two new decision traps: the problem of over-confidence and the puzzling phenomenon of sunk costs. Philosophers have also discussed the human tendency to fail to decide and fail to act. The existential philosopher Walter Kaufmann talks about decidophobia and Aristotle wrote about akrasia - weakness of will.

Applications: Along the way we'll emphasize some classic applications of RCT. The distinction that statisticians call Type I /Type II error and medical researchers call false positives vs. false negatives can be applied in all sorts of contexts, including an analysis of the way our judicial system is supposed to work.
We'll also see what RCT has to say about a few literary examples, such as the stories of Ulysses and the sirens and the wisdom of Solomon in a child custody case.

Some of the material in this unit could have been presented in a superficial way at the beginning of this course as a way of motivating interest in the subject. But I decided (there's that word again!) to use it now instead, partly as a review, but mainly to show how one can move from a rather superficial, common sense appreciation of these issues to a deeper understanding that arises out of familiarity with both the strengths and the limits of Rational Choice Theory. Aristotle provided a sophisticated account of what he meant when he characterized humans as Rational Animals, but I think this course clearly demonstrates that we now have a better understanding of both the rational and the animal aspects of human nature.