By Briana Wipf
About 20 attendees gathered on Zoom to discuss the ethics of using student data for learning analytics on Thursday, Feb. 10, in a workshop led by George Rehrey of Indiana University in Bloomington.
The workshop, sponsored by the Year of Data and Society, is one of multiple events focusing on student data. The second is a faculty panel, also on Zoom, on Wednesday, Feb. 23, from 3-4 p.m. The third session, also on Zoom, is on Thursday, March 17, from 4-5 p.m. and is targeted at students.
Rehrey’s presentation and following workshop, which was co-facilitated by Lizette Muñoz Rojas, a teaching consultant at the university’s Center for Teaching and Learning, occurred at a time when the discussion around the collection of student data has become more relevant than ever.
Rehrey is the founding director of Indiana University’s Center for Learning Analytics and Student Success, which “advances the widespread use of big data and student learning analytics, empowering faculty to conduct actionable scholarly research through the systematic collection, exploration, and analysis of data describing students, their observable activities, and outcome,” according to its website.
Rehrey said that he often uses the metaphor of the “wild, wild west” or “a train leaving the station” to describe how student data is collected and used, and the policies around that collection and use.
“We’re bumping up against this whole new frontier,” he said, relying on another metaphor to help attendees conceptualize both the novelty of the issue and the high stakes.
When the term “student data” is used, it may refer to more than one type of information. Some data is strictly administrative and includes information about demographics, enrollment, application data, and other categories. At Pitt, this administrative data is collected on PeopleSoft. This where students’ official records of credits earned and grades is kept and visible to Pitt’s administration and staff who keep track of degree completion and graduation eligibility, according to Amanda Brodish, Pitt’s Associate Vice Provost for Data Analytics, whom I interviewed following the workshop.
But de-identified data from PeopleSoft may also be available to faculty researchers who want “to conduct educational research that will contribute to what is known about teaching and learning,” according to the Student Data webpage. Students are able to opt out of their de-identified information being included in such data, Brodish told me.
Another type of data is that which instructors might look at related to students’ performance in their classes, and this type of data is available on a Canvas class site. Only the course’s instructor can see this type of data, and an instructor can only see a student’s data for that one course.
“We wouldn’t be sharing Canvas information in the same way that we might be sharing academic information in PeopleSoft,” Brodish said.
Rehrey pointed to a 2021 Inside Higher Ed story written by Melissa Ezarik that shared results of student surveys regarding data collection. Only 12 percent of students surveyed knew about their institution’s data collection policy and had read it, for example. Of the remainder, 37 percent of students were aware a policy existed but had not read it, and 36 percent were not aware that such a policy existed.
The survey also found that students said they had “no concerns” about the collection data on attendance (66 percent), grades (55 percent), or enrollment (51 percent). Indeed, 85 percent of students approved of push reminders for assignment due dates.
Canvas, Pitt’s learning management system, has a data analytics function that can be accessed by the instructor of a given class. The “New Analytics” functionality allows instructors to see information about students’ grades and activity on the class site.
Much of Rehrey’s presentation and the workshop component, which saw groups of attendees discussing the ethical concerns of two hypothetical scenarios, dwelt on the tension between using data to help a student, and of relying on data points – of grades or time spent on the class site, for example – with little additional context.
“Be careful with what I like to call unintended consequences of the use of this data,” Rehrey cautioned the group.
For more on student data and privacy, Brodish recommended two webpages on the Registrar’s website, one for faculty and staff, and another for students. Students can opt out of having their data shared or used in various ways, including their directory information being shared with a third party; their administrative data for scholarly research purposes; and results of predictive modeling being used with advisors and administrators.