How to Make the GPA Less Flawed

Whether we like it or not, GPAs matter. For decades, American society has reduced student success, or failure, to this number completely absent of any context. Colleges and universities themselves use grade point averages in many important ways, including determining which students receive scholarships and honors.

Although more and more employers and graduate schools proclaim to place little emphasis on an applicant’s GPA, they often use it as a first-pass filter to weed out supposedly weaker applicants. According to the Job Outlook 2018 survey, conducted by the National Association of Colleges and Employers, more than two-thirds of employers plan to screen the Class of 2018 candidates by GPA.

Dan Black, global recruiting leader at Ernst and Young, noted in 2014, “There is no hard-and-fast rule, but the lower the GPA, the more evidence we need to see of other competencies … If someone has a GPA in the lower 3.0 range, we would want to see that they had a part-time job to help get through school, or played a sport, or were active in student government. There needs to be a reason [for it].”

GPA has acquired such an important role because it seems to be a convenient statistic to quickly compare students, either to each other or to benchmarks of excellence. However, using the GPA in that way assumes that grades are awarded consistently, which couldn’t be further from the truth. Grading norms vary significantly between courses and instructors, even within the same institution.

It is hard to tease out all of the reasons that grading norms vary so much. Variations in grading norms are, in fact, sometimes necessary depending on the course and teaching methods used. Consequently, imposing uniform grading norms across courses is not the answer. For a decade, Princeton University tried a universitywide imposition of uniform grading “targets” but abandoned the policy in 2014.

Bucknell University’s Committee on Instruction, while not interested in repeating Princeton’s experiment, nevertheless became concerned about the effect of grading variability on the fairness of GPAs. It conducted a study of the GPA of all of the graduating students from the Class of 2016. In addition to GPA, the committee also calculated the “GPA of Medians,” or GPAM, for each student -- a metric that reveals the GPA a student would have received had they obtained the median grade in all their classes.

The results were stunning. The range of GPAMs varied from a low of 2.91 to a high of 3.85. This means that some students consistently took courses that had median grades of B or below, while other students in the same graduating class consistently took courses with median grades of A or A-minus. As a result, comparing the GPAs of students can be misleading.

Consider the following five students from the 2016 study, first viewed solely through the lens of GPA:

Based on GPA alone, Student No. 5 would appear to be the strongest academically, whereas Student 1 would appear significantly less successful.

Now, consider the same five students with the added information that GPAM provides:

The story seems quite different. Whereas GPA alone implies that Student No. 5 was much more successful academically than Student No. 1, the GPAM reveals that both of these students did work that was comparable to the medians in their classes. Using GPA alone would also imply that Student No. 3 outperformed Student No. 2, but GPAM shows that Student No. 2 significantly outperformed peers in the same courses whereas Student No. 3 consistently earned grades below the median. The GPAM information also reveals that Student No. 4 not only graduated with a high GPA, but achieved that GPA in courses that typically give relatively low grades.

Given this data, one wonders if the GPAM could help, say, Student No. 2 make a compelling case to Ernst and Young that their GPA is evidence of strong academic performance. It does not need to be “explained” by participation in a sport or student government.

Following the study, Bucknell’s Committee on Instruction proposed the addition of the cumulative GPAM to academic progress reports. When presented to the Bucknell Student Government, the response was overwhelmingly positive. In fact, several students were eager to see their own GPAM. In spring of 2018, Bucknell’s faculty voted overwhelmingly in favor of adding GPAM to academic progress reports. In March 2019, the statistic became available to eligible students for the first time.

Bucknell is not the first university to consider providing information about grading norms as a way to address the issues caused by grading variability, but the GPAM approach seems to be distinct.

Cornell University and Dartmouth College both implemented policies where median grades for every course are included on transcripts. Concerns have been raised that this approach can expose individual faculty members to potential scrutiny for pedagogically justified grading practices. Furthermore, at Cornell, that practice was linked to accelerated grade inflation, as students were using the information to shop for courses with higher grading norms.

In 2014, the University of North Carolina at Chapel Hill voted to add median grades to transcripts, along with a cumulative, average grade statistic called Schedule Point Average, but as far as we know neither the median grades nor that statistic ever actually appeared on transcripts.

From a faculty perspective, GPAM addresses the inequities presented by GPA without dictating to faculty members how to do their job. And while GPAM doesn’t solve all of the problems with GPA, it does make a seriously flawed measure less flawed.

If more people outside Bucknell become aware of GPAM -- and other colleges and universities adopt the measure -- employers might start asking to see a student’s GPAM. In that situation, having a lower GPAM may signal that a student is willing to take risks to pursue any academic path.

Despite its obvious flaws, the convenience of a single number makes it unlikely that GPA will disappear from higher education. By providing some context to an individual’s GPA, we allow students to better understand and market their accomplishments.