We designed a rubric to assess free-response exam problems in order to compare thinking skills evidenced in exams in classes taught by different pedagogies. The rubric was designed based on Bloom's taxonomy and then used to code exam problems. We have analyzed historical and recent exam problems in both algebra-based and calculus-based exams. In particular, we have examined cases where the same problem was administered on exams taught by different pedagogies. In some cases, we were able to compare different sections of the algebra-based physics course taught by the same instructor, sometimes during the same semester, one with inquiry-based instruction and the other in a more traditional lecture environment. We discuss the development of the instrument and present results. The inquiry-based students consistently demonstrated use of the thinking skills coded at least as often as the traditional students, and usually more often.