The testing dilemma
Many of the students in our biology classes want to be doctors, and I recently stumbled on what looks like a major effort to update the pre-med curriculum, sponsored by the medical colleges association and HHMI. The thrust of the report is that to get into medical school, students should show “competency” around eight different areas of science (there are also an additional eight science competencies for medical students themselves). It’s an interesting break with transcript-based judgement of student achievement, and aligns with a lot of current thinking about science education.
The core competencies the authors came up with are quite reasonable and would probably get near universal agreement for importance. They are things like “Apply quantitative reasoning and appropriate mathematics to describe or explain phenomena in the natural world.”, or “Explain how organisms sense and control their internal environment and how they respond to external change.” Under each of these broad categories are learning objectives which in some cases seem a bit of a laundry list, but again most of us would probably agree they are generally good goals. The whole document looks like a well-thought out attempt to define what future doctors should know how to do.
Underlying this switch from passing particular courses to showing particular competencies is in some respects a switch from knowing something to knowing how to do something. A competency, such as applying quantitative reasoning to biological problems, is different than mastering an isolated body of knowledge, such as having passed a calculus class and remembering a bunch of integrals. Not surprisingly, I like that, since helping students learn ideas in a way that they can use them drives much of our design of virtual labs and interactive chapters at SimBio. But I predict that implementing this change for medical school admission will run into a pretty high wall.
Depending on competency rather than course grades puts an even greater burden on a bugaboo of education – testing, and especially standardized testing. Medical schools rely on standardized tests to help them sort through all the students who want to be doctors. It’s a lot of students (it often seems like just about all the students in a typical intro bio class), so some sort of standardized test like the MCAT is here to stay for admissions. Right now, we don’t really have good standardized tests to distinguish whether a student simply remembers lots of facts, and has figured out where to plug each one in on an exam, or whether the student actually understand concepts and can use them to think through problems. We basically hope that the first predicts the second. But I don’t think that will be enough to change how pre-med classes are run across the country. Somehow the tests have to really pick up, directly, whether the student is competent enough with each set of ideas to use them effectively in thinking about medical topics.
We don’t, by and large, have those tests now, but we have some clues about how to make those tests. Using concept inventories and careful qualitative research on student thinking, we can design better test questions for specific topics. But more broadly, computers have the potential to allow standardized tests that make students apply their knowledge by solving problems. With some help from NSF, we did a project this past spring on designing one such test for technology literacy (we’re calling it SimBio Testing Tools), and other groups are also making such attempts. I think this is a great opportunity. Reports like this one on medical students are piling up, so biology education is ripe for change. If we could just measure it, that could be the tipping point.