I just returned from my annual sojourn teaching at the ICPSR's Summer Program in Quantitative Methods. Most of the courses there -- which range in technical difficulty from Introduction to Computing and basic math to Bayesian methods, LISREL, and advanced game theory -- are four-week courses, populated mostly by graduate students from Ph.D. programs in sociology, political science, and other similar fields.
One thing I've always wondered, though, was how the program might attract more individuals who are later in their careers. A tenure-track professor in any discipline can hardly afford to give up a month of their life to learn a statistical technique they may only use for a single project. Yet, it's often exactly such individuals who could most benefit from training of this sort.
The ICPSR has begun moving in this direction in recent years, offering greater numbers of three- and five-day workshops and holding courses outside of Ann Arbor (which is, however pleasant in the summertime, a bit off the beaten path). But they (we) could obviously do more. So I ask: What would you look for in a summer (or winter/spring break) course in methods? What characteristics (pedagogical, logistical, whatever) would make you more likely to enroll in such a course?