College students, according to data specialist Adam Lang, are like fish. Observing the movement patterns of fish and the academic practices of students is roughly the same, conceptually.
Confused? It’s okay, so was I.
The New York Times has reported on the opportunities of data mining in higher education, focusing heavily on the country’s largest public university, Arizona State. Due to its massive population of over 72,000 students, the university is popular for data-driven experiments. The school has implemented a degree-monitoring system that keeps track of students’ performances and raises flags when one strays off course.
While Lang is comparing college students to fish, we might as well compare ASU’s system to Netflix. Netflix is able to recommend movies and shows based on your previous viewings and your participation in rating films and TV shows on a five-star scale. In a similar way, an algorithm is able to observe what classes a student took and how they performed to recommend further classes that student might be interested in taking. Additionally, the system can warn students by web alerts if they are off track in a given degree program based on their rate of assignment completion, test scores and other factors.
Sounds kind of Orwellian, doesn’t it? As a school of 72,000, it stands as no surprise that ASU is embracing technology to handle its titanic empire. But what does this mean for the role of professors and the traditional college model?
ASU’s data-mining efforts are driven by a perceived need to reduce dropouts and improve graduation rates. By working in the background and using data to guide students with email alerts and “eAdvising,” the university feels it can affect these numbers. But at what point will this begin to feel like a diploma mill, a soulless experience that rolls students along a production line? The article reports that “Tennessee, for example, doles out higher education dollars in part by measuring how effective an institution is at graduating students.” This is missing the point entirely. Many college professors would argue that the point of higher education is not to churn out students like cars through a factory line.
Some would also caution against treating students like nothing more than sources of data. “I’m worried that we’re taking both the richness and the serendipitous aspect of courses and professors and majors—and all the things that are supposed to be university life—and instead translating it into 18 variables that spit out, ‘This is your best fit. So go over here,’ ” said Michael Zimmer, assistant professor in the School of Information Studies at the University of Wisconsin, Milwaukee.
Still waiting for that fish analogy to be explained? Adam Lang is a computer science major who now works for Ellucian, a higher-education company that provides data-mining software to universities. A data-fiend and tech guru, Lang monitored his fish tank and observed the swim patterns of his Betta fish. At the same time, while he was in college, Lang became intrigued by the behavioral data of his fellow students. Lang and his student colleagues measured the data left behind by other undergrads, taking note of log-in rates, at what point in the semester they—finally—viewed the syllabus, when homework was turned in, etc. From all of this he theorized that one could determine pass/fail rates of students, equating the predictability of fish movements to students’ academic behaviors.
Lang and Ellucian’s software can help a professor identify students who may need extra help, but it cannot and should not stop there. As with any new educational technology, professors must never use it as a substitute for traditional student/instructor mentoring. There is nothing wrong with mining data and observing patterns from students. After all, colleges are sub-cultures in themselves, and as an anthropologist might gain knowledge from observing the habits of a culture, statisticians can gain knowledge from observing the habits of students and put that information to good use. But institutes of higher learning must be cautious of thinking of their students as lab rats. Universities can use data mining to guide students along the right path, but never at the expense of personalized and quality education, the most important element being student-to-teacher interaction.