Knowing previous students’ grades affects undergraduates’ GPA, Stanford scholars find
Undergraduate students who used a course information platform to learn about the performance of previous students in classes at their university earned slightly lower grades, according to new Stanford research.
When undergraduates know how previous students performed in courses at their university, they end up receiving slightly lower class grades on average than they would have if they did not check out that information, according to Stanford research.
In a new research paper, an interdisciplinary group of Stanford scholars examined the effects of a course-planning web application which visualizes data from registrar records and prior student evaluations for each class. Through a randomized field experiment, the scholars found that the use of this platform led on average to a drop of 0.16 units in overall GPA. As an example, that decrease is large enough to move a B+ grade roughly half the distance to a B, the researchers said.
“The information students rely on to make academic decisions can have a demonstrable effect on their behavior,” said Mitchell Stevens, an associate professor at the Stanford Graduate School of Education and a co-author of the paper. “The design of academic environments, both digital and physical, really matters for student outcomes in college.”
The lowered scores were especially prominent among freshmen and sophomores, whose GPAs declined by 0.26. Juniors’ and seniors’ GPAs decreased only by 0.09 on average, according to the researchers.
The findings offer some of the first evidence of student behavior and academic performance being affected by exposure to this type of data, used more and more often in higher education as a way to help all students have equal access to information about courses and make more knowledgeable decisions about their academic schedules.
“This is a powerful insight for educators, because it suggests that the presentation of currently available institutional information can influence students’ academic behaviors,” the researchers write.
“The main takeaway here is that providing this type of information in a university environment can have a wide range of effects, some of which are unexpected. And we need to understand more about how students use some of this information,” said Ramesh Johari, associate professor of management science and engineering and the study co-author.
The new research paper is scheduled to be presented at the ACM Conference on Learning at Scale at the end of June. Tum Chaturapruek, a computer science doctoral candidate, is the lead author on the paper. Additional authors on the paper are Thomas Dee, a professor of education and director of the Stanford Center for Education Policy Analysis, and René Kizilcec, a researcher at the Graduate School of Education.
Behavior and information
Universities and colleges around the country are starting to provide students access to data to help them make decisions about the courses they take.
For this study, the researchers examined the impact of students’ use of Carta, a web-based course-planning tool developed by the researchers and made available to students as a voluntary resource to help them plan their courses. The students were divided into two groups: those who were encouraged to use Carta and those who weren’t.
The Carta platform provides course information, when available, including the distribution of prior students’ grades, percentage of students who dropped out and withdrew, average rating given by previous student evaluations and the number of hours per week students reported spending on the course, plus advice from past students to future students considering taking the class.
“Universities have a treasure trove of data on what students have done in the past, and this data can help students make better choices,” said Kizilcec, who is joining Cornell University as an assistant professor of information science in July. “One of the big questions is how to present this data so it could guide them rather than lead them astray.”
“Students start college with a lot of uncertainty about how hard their coursework is going to be. So when they see those grade distributions, which often show that a high percentage of students get A’s, they may invest a little less effort in their coursework.”
Associate Professor, Stanford Graduate School of Education
A concern some educators have expressed about showing students previous grades is that it could lead to students choosing what look like easier courses in order to get higher grades, resulting in grade inflation.
“However, the evidence from the study does not support the conclusion that students change their course portfolio to take easier courses after using Carta,” Chaturapruek said.
The researchers found that seeing prior grades had the biggest impact on GPA. But the researchers also found evidence that showing students solely how much time was spent on the course had a positive effect on GPA.
“The reality is much more complicated,” Johari said. “Students use this information in different ways, which we are just starting to try to understand.”
What might lead to the grade drop? The researchers found that the shift in GPA was tied to changes in students’ behavior within courses. The researchers hypothesize that the GPA changes are tied to students’ expectations about how hard their courses will be.
“Students start college with a lot of uncertainty about how hard their coursework is going to be. So when they see those grade distributions, which often show that a high percentage of students get A’s, they may invest a little less effort in their coursework,” Stevens offered as a plausible interpretation of the finding. “This is substantiated by the fact that the negative GPA change is strongest for first-year students during their first quarter.”
Beginning point for discussions
The new research paper is the first in a series of studies the researchers are conducting. They also hope to partner with other universities on future studies.
Further research is needed to figure out how the behavior of students changes exactly, the scholars said. In addition, the effects of similar information on student behavior may be different at other universities or colleges.
“This research is already creating many conversations, and there are great questions out there: What sort of information do students deserve to know? In what form should that information be presented?” Stevens said. “We hope to continue to contribute to these discussions and others with further studies.”