I wanted to respond to Jacob Levy's post on research vs. teaching, and then one of my teaching assistants asked me about the same topic over beer at the end of the semester, Steve-thanks-his-TAs-with-beer festivities. Jacob linked to an article that asserted that teaching does not matter much for promotion/merit pay by universities, and Jacob appropriately notes that this falls under the category of uber-obviousness (my term, not his). Perhaps at liberal arts colleges, teaching comes into play, but not so much at research universities.
In my experience, just as in this article, pay and promotion hang entirely on one's research output. When I have served on tenure committees, teaching was largely an afterthought. The only time I have seen teaching matter was for a brief period of time, where the chair had created an entirely transparent point system to provide clear incentives/rewards for publishing in the most visible journals, doing well on teaching evaluations, and providing good service to the department. He was, by the far, the best chair I have had thus far. His system was, perhaps obliquely, aimed at rewarding those who were most likely to leave, at the expense of less productive folks. As a result, some people were mighty unhappy. And, strangely enough, he had the shortest term of all the chairs I have had.
Of course, if one cares about teaching, that is just the first step. Teaching effectiveness is hard to measure, and professors have a real hard time evaluating each other anyway, even when it comes to counting how many articles and books someone has. Good teaching evals and even teaching awards can mean someone is a dynamic lecturer, easy grader, or both. Or neither. Some schools do serious peer evaluations,* but, just as professors are not trained to be administrators and have largely not been trained to be teachers, they have not been trained to evaluate teaching either. On the other hand, research effectiveness is also hard to evaluate (see my previous posts on co-authorship and reputation), so perhaps I should start over.
Is there a tradeoff between teaching and research? Absolutely, as both require time, and the more time puts into one, the less there is for the other. But, that does not mean that the tradeoff is absolute. That is, time invested in one can contribute to the other. Teaching can inform research, and certainly research informs teaching (only service seems to be an absolute time sink with little positive impact on the other two parts of the job--or am I over-stating?). I have developed new research questions and sharpened my thinking through the act of presenting the material to my students and through their questions. Indeed, I have taught new classes the past few years to develop a better grasp of new areas in which I want to research. And I have definitely brought my research into the classroom.
Providing solid incentive structures is hard enough in the business world where we have seen individuals and firms develop perverse incentives that lead to them to short-term riches and long-term catastrophes. When it comes to harder things to measure than revenue streams like quality teaching and significant research contributions, we should not be surprised that colleges and universities have a hard time and get it wrong.
* Peer evaluation is kind of like the cops evaluating whether the other cops are not abusing their authority. That is, people tend to protect their kin. At most universities, department decisions are only the first step, as those closest to the person being considered for promotion may be inclined one way or another regardless of the of the merits but due to something else. So, the additional layers of review are aimed at taking a more objective look at the record.