Data Mining to Improve Education

On July 18, 2012, the CHE published a lengthy piece called “College Degrees, Designed by the Numbers.” Author Marc Parry presents a variety of ways that Arizona State is using behind-the-scenes browsing data to customize the learning experience:

  • math courses that direct students to additional practice and tutoring when they struggle
  • Facebook apps that suggest friends with common interests
  • online data collection that predicts — in just over a week after the semester begins — which students are in danger of failing an online class

Arizona State even requires students to change their majors if too many red flags (dropped classes, low grades in gateway courses, failure to enroll in classes in the major) pop up. As Parry notes, “some students show up with ambitions that bear no relation to their skills. Or parents push them into majors that don’t interest them. Or they feel like shoppers in a cereal aisle, confounded by the choices.” The goals are to push students to the majors where they can shine, to decrease time to graduation, and to increase graduation rates.

On the one hand, I think this is a good idea. Kids who barely got through algebra arrive to college planning to major in engineering, completely unaware of how much math is involved. While some of these students will rise to the challenge, study until their eyes blur, and master calculus, others will struggle for two years and end up with a pile of credits that don’t count towards another major.

Additionally, with parents and government officials wanting quicker graduation times, students don’t have the luxury of shopping around — to walk across that stage in four years, students need to complete 30 semester credit hours (approximately 10 typical classes) each year. High school doesn’t always provide much opportunity to check out different careers, either, as more districts move to a highly academic, college prep curriculum in English, math, science, and social studies. Fields like art, construction, social work, psychology, agriculture, engineering, criminal justice, and business are electives, if they are offered at all. Sometimes there’s a stigma attached to such courses, i.e. “you’re taking Principles of Nutrition instead of AP Chemistry? oh, you must not be very…you know…smart.” We expect American teenagers, at the age of 18, to know what they want to study for the next four years when all they’ve done is take what they’ve been told to take — maybe it IS best to keep doing some of that for grades 13 to 16.

Still, that aspect of data mining doesn’t excite me as much as another project that Parry mentions: using data to suggest courses at registration time.

I advise students. I advise a lot of students. Some come to advising sessions with battered degree plans in one hand and a list of all courses (organized by time slot, with alternates for everything in case classes aren’t available) in the other. The other 96% come to my office and ask plaintively, “so what do I need?!” They generally look like this:

student looking stressed and confused

Creative Commons licensed content from CollegeDegrees360

Of course I make plenty of suggestions, based on the degree plan, but what if there was another layer of advice? Austin Peay State University in Tennessee uses Tristan Denley’s software, which “melds each student’s transcript with thousands of past students’ grades and standardized test scores to make suggestions. When students log into the online portal, they see 10 ‘Course Suggestions for You,’ ranked on a five-star scale….the software bumps up courses for which a student might have a talent, by mining their records — grades, high-school gradepoint average, ACT scores — and those of others who walked this path before.”

That’s something I can’t do in any scientific fashion. I can recommend a core humanities class that I know hundreds of other students have passed, but maybe the advisee in front of me has talents better suited to something else. Because the course portal simply suggests 10 classes, students aren’t obligated to follow any of the computer’s advice. When I asked my students via Twitter if they would like having this kind of option, suggestions for classes to take, they said yes.

There are many reasons for that answer, but Parry’s article gives an explanation that rings true, at least from my experience:

When presented with many options and little information, people find it difficult to make wise choices. The same goes for college students trying to construct a schedule, he says. They know they must take a social-science class, but they don’t know the implications of taking political science versus psychology versus economics. They choose on the basis of course descriptions or to avoid having to wake up for an 8 a.m. class on Monday.

One person interviewed for the article laments that this kind of data mining removes the  serendipity of the college experience, the stumbling across a subject you never considered before. But today’s students? They want to finish. Serendipity isn’t as important as financial stability and a diploma in hand.

So put me down for a beta test of “Your 10 Top Course Suggestions” if a Texas university is needed!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s