First, you need to read good books. To get the most out of those books, you need to talk about them with other people who are also trying to work their way through them. In addition, you need to write about them in a disciplined and focused way. Both of these tasks require supervision and guidance by more experienced learners — preferably those who have already gone through an educational program that takes both discussion and written analysis to the highest level.
Second, for some types of skills — such as language acquisition, mathematical manipulation, and technical lab skills — there’s no way around requiring carefully targetted and supervised exercises. Preferably, these exercises would be developed and overseen by someone with a high degree of technical proficiency and experience in the field in question, as such a person would have the best view of which skills were most valuable.
Finally, for command of facts, limited use of rote memorization can provide a baseline, but the main focus should be on learning how best to search for information and assess the trustworthiness of the sources found. All of this is best done in close dialogue with someone who has a lot of experience with research.
I believe that the pedagogical research would bear all this out, and my own experience at an institution that embraces this model shows me that it works.
Starting from these premises, certain natural consequences inevitably present themselves.
First, we must minimize the number of full-time faculty at our institutions and rely heavily on less experienced instructors such as grad students. Second, we need to invest as many resources as possible in administration and capital-intensive projects such as new buildings and cutting-edge computer technology. Third, in order to increase access to quality education, we need to vastly “scale up” lecture courses to the point where students will have no direct contact with the instructor at all. (Indeed, we should repeatedly re-use recorded lectures so that the instructor doesn’t even need to be involved with the course on an ongoing basis.) Finally, in developing our programs, we need to be guided above all by the whims of wealthy donors — after all, if they’re willing to pay for it, it must be worth doing.
Only by using such an evidence-based model of higher education can we continue the tradition of excellence that has made the American university system the envy of the world.