GPA Does Not Forecast Who Finishes a PhD and How Quick

Students with a low and high GPA during their undergraduate studies fared similarly in their doctoral studies, suggesting that GPA is not an accurate predictor of research success.

As a biology professor at San Francisco State University (SFSU), a minority-serving institution, Megumi Fuse often encountered students from under-represented communities. She observed that since these students frequently had to balance their studies with work and other familial responsibilities, they could not divert all their focus on academics, which would affect their grade point averages (GPAs).

However, she realized that the students’ low grades were not necessarily an indicator of their research aptitudes. “I had this experience that so many students with a low GPA were successful [in research],” said Fuse. Despite this observation, she had no way to quantify this at the time.

She decided to follow up on the observation when she took over as the director of the Student Enrichment Opportunities office, which provides academic and financial support to prepare students for STEM careers. The department had collected long-term student data that she could examine.

By analyzing data over 30 years, Fuse and her colleagues found that a high GPA during undergraduate studies did not predict greater success in entering or completing a PhD, or how long it took to earn the degree.¹ Their findings, published in PLoS One, indicate the limitations of applying strict GPA cutoffs for applications to PhD programs and the need for broader criteria for admissions.

“The paper is quite convincing,” said Jasjeet Singh Bagla, an astrophysicist and educator at the Indian Institute of Science Education and Research Mohali. While scientists have anecdotal observations, “the nice thing about the paper is that it does hard statistics and clearly shows that undergraduate marks…don’t predict how well a person will do in terms of research.”

Fuse and her colleagues investigated whether undergraduate GPA had an association with success in doctoral programs for students who were mentored for research. They assessed the records of students from 1992 to 2019 who participated in a research-focused Master’s degree program at either SFSU or California State University, Los Angeles (CSULA). These students had an undergraduate GPA between 2.0 to 4.0.

Of the 506 students they tracked, nearly 88 percent entered PhD programs. This proportion did not differ significantly between those with a GPA higher or lower than 3.0.

Next, the researchers examined whether a higher GPA meant that the students had a better chance of completing a PhD. A comparable number of SFSU Master’s students with low and high undergraduate GPAs went on to complete their PhDs. In contrast, a greater percentage of CSULA students who finished their PhDs came from the low undergraduate GPA group.

Fuse and her team further observed that among those who completed their doctoral studies, students with both low and high GPAs took a similar amount of time.

While Fuse expected that GPA would not be linked to PhD success, she was surprised to see how many students actually ended up entering a PhD program after completing their Master’s. “It speaks to this need for students to have mentorship and a guided program to bring out their abilities,” she said.

Bagla agreed that the absence of a link between undergraduate GPA and PhD success is not entirely surprising. “In undergraduate studies, you are doing courses. You are solving problems which your instructor knows can be solved,” he said. “In research you are working on problems which no one has solved before. Therefore, it requires a different kind of a mindset.”

However, he noted that it would have been interesting to see the research status and outputs of some of the older students. He added that the researchers have assessed a limited number of students from limited backgrounds, but he believes the results will hold true across different cohorts. These results emphasize the need to look beyond just grades in the field of research, he said.

Fuse agreed. “The overarching implications are that there has to be a different way of assessing applicants that doesn’t go to an old school model.”