I heartily recommend Leonard Cassuto’s recent article in the Chronicle on graduate school completion rates. If we admit to our graduate programs only those students who are dead certain about their occupational direction, and whose academic record affords us equal certainty about their success, we will be making at least two mistakes:

1. We will be taking no risks. We will miss the chemistry major who figured out late in her college career that she would rather be a historian, and whose profile therefore looked a bit odd and included a GPA pulled slightly downward by her demanding courseload.  We miss the naïf who, at age 18, selected a school based on an occupational assumption that couldn’t survive the attraction of history books assigned in an elective but who couldn’t transfer. We miss the working-class student who attended a local non-elite public institution and hence lacked the profile of the polished graduates of elite liberal arts colleges. We miss the first-generation college student who had no idea that one could even make a living doing history, but who had a professor who told her that she was talented and should give it a try. All of these examples are of real people who had “non-traditional” starts (and hence not “dead certain” admission status), but still attained professional success as historians and contributed in important and diverse ways to the discipline.

2. We will be embracing false certainties. The best students are people who learn.  And people who learn often figure out that a decision made on the basis of incomplete knowledge is not the right decision. So, if we want “the best students,” then we should be open to those who are most likely to reflect thoughtfully and critically on their graduate school experience—which means that many are likely to change their minds about their future careers. Besides, our certainties about who is likely to be successful based on academic data are highly problematic.  Why should we expect our predictive powers to be any better than professional sports franchises that invest zillions of dollars into developing and running tests on their prospective draft picks?  Why should we assume that our grad school admissions predictions be substantially more accurate than top draft choices?

This post first appeared on AHA Today.