I am at Washington University for the Advanced Course in Empirical Legal Scholarship. For the last few years, Lee Epstein and Andrew Martin have offered a basic course (see here for an earlier endorsement). This is the first advanced course they have offered, and I think it is a tremendously time and cost effective way to cover approximately 95 percent of the skills needed to conduct and interpret empirical scholarship. Lee and Andrew are extremely well prepared instructors who offer a very good mix of lecture and class exercises. Their use of PowerPoint also made a big difference in organizing and simplifying the material.
The topics covered so far include the basic assumptions of linear multivariate regression, including diagnostic methods and corrective measures; effective presentation of regression results; application of the Monte Carlo principle; a fairly in-depth treatment of probit and logit methods for ordinal and nominal dependent variables, including interpretation; and a fairly brief but useful explanation of panel data, time series, survival analysis, and several other methods. Tomorrow we are going to cover nonparametric matching techniques, which offer (I am told) a better way to deal with endogeniety than instrumental variables in a linear model.