The University of Arizona recently announced its Smart Campus research initiative uses data collected from students’ ID to predict if a student will drop out or transfer to another institution. By tracking freshmen students’ CatCard swipes from almost 700 locations on campus, the University hopes to learn how to better retain first-year students.
Sudha Ram, UA professor of management information systems and research lead, claims these “digital traces” tell the story of a student’s campus and social life experience, which may correlate to their likelihood to leave the University. Ram theorizes student social behavior correlates to retention rates: the more a student is involved and have friends on campus, the more it is likely they will stay at that institution. After collecting three years’ worth of data, and through the use of machine learning, Ram and her team boast a 85-90% accuracy rate when estimating which first-year students won’t return to the University of Arizona for their sophomore year. Impressive, to say the least.
Student retention, especially first-year to second year matriculation, is a long standing issue in American higher education, an issue that many campuses are hoping to chip away at using big data. Georgia State currently analyses “800 risk factors” associated with attrition rates. A number tech startups have popped up to sell campuses products using predictive analytics to improve student success (aka retention). In mid-2017 the National Association of Student Personnel Administrators (NASPA) published a guide it to help administrators use big data to promote degree persistence (aka retention, too). Collecting all this data can’t be bad if the goal is to keep students in school, right?
Well, it depends. How does the institution collect the data? How will it use the data? Do students know they are being tracked? Have students consented to participate in the study? FERPA, the Family Educational Rights and Privacy Act of 1974, does not contemplate big data, or any digital data for that matter, not relating directly to the student’s academic performance, personally identifiable information, or student conduct record. FERPA does includes provisions protecting information if it is used for research as long as it is de-identified if published. Thus, institutions and third party vendors collecting big data aren’t doing anything illegal, but their actions may be ethically dubious.
Never before have colleges and universities been able to literally track their students. NASPA’s guide referenced two universities planning to use GPS and their campus Wi-Fi network to send students “notifications or text messages encouraging [them] to use the services or attend events that are nearby.” Colorado State uses big data to create support interventions when a student’s behavior triggers their early warning system. Reports indicate that students suspect their college or university tracks them in some capacity, but is it the institute’s business every time a student goes to the campus convenience store to buy condoms?
Is all this big data even painting an accurate picture? UA’s Smart Campus algorithm determines that two students have “social interactions with each other” if those two students demonstrate a pattern of being in the same place at the same time. At a university with 45,000 students, it’s likely many students have the same daily habits and patterns as their total-stranger peers. Ram admits the data collected from ID swipes are “not designed to track [student] social interactions, but you can, because you have a timestamp and location information.” Secondary data analysis is one thing but using data in your study that was not originally collected for the purposes of research isn’t great methodology.
If all of this data is used to assess the factors that contribute to student persistence, i.e. grades, degree planning, advising, financial aid, factors of which are protected by FERPA, should not that information be protected as well? As of right now, this data belongs to the collecting institution, leaving students without the right to their own movements and actions on campus. Is it time for big data become one big student record?
Can a Middle-Man Actually Reduce Costs?