The Washington PostDemocracy Dies in Darkness

Big data could solve the college-dropout problem

New York Institute of Technology’s Old Westbury campus (Photo by Christopher Appoldt.)

Mark C. Hampton is vice president for planning, analytics and decision support at New York Institute of Technology.

America’s most famous college dropout thinks too many students are following his lead.

Bill Gates, who left Harvard to found Microsoft, says it’s “tragic” that only half of the 2 million students who started college this fall will graduate.

In a recent blog post, Gates hints at a solution: big data.

Colleges have access to extraordinary amounts of data — not just on students’ academic performance, but also on their eating habits, social life and daily routines. Administrators could use such data to identify which students are at risk of dropping out and could then intervene to give them additional support.

Some institutions already do this, as Gates pointed out: Georgia State University examined more than 10 years of transcripts from 140,000 students. GSU identified 800 risk factors that make it more likely a student will flunk out. It found, for instance, that only 1 in 4 political science majors who earned a C in his first class graduated on time.

Now, the school sets up counseling sessions with students at risk of falling behind. In just a decade, GSU cut dropout rates by 32 percent.

GSU’s system is remarkably effective — but it still doesn’t harness the full potential of student information, because it only examined academic records.

Imagine the insights colleges could glean from a much broader array of data about students’ lives. Schools gather data every time a student swipes into a dining hall, exercises at the fitness center, checks out library books and joins a club.

All this data could help identify struggling students before they drop out.

Consider a student who frequents the campus gym and participates regularly in student activities: If this pattern abruptly stops, the school could check in to make sure everything is okay.

Unfortunately, most colleges don’t analyze this data in real time. They often wait until students leave school to take a look at the stats.

That’s a mistake.

Private and nonprofit organizations routinely use big data to better serve clients and customers. Netflix analyzes viewing histories to predict which movies subscribers may want to watch next.  UNICEF crunches health survey data to determine which communities have the greatest need for development aid.

Some might worry about colleges collecting and analyzing these kinds of data about students.

But there’s no cause for alarm. Students already give information about their social activities and locations to companies such as Facebook and Google — which use that data for profit.

98 personal data points that Facebook uses to target ads to you

Colleges, at least, would use the data to improve students’ lives.

Nevertheless, colleges and universities could impose some restrictions on themselves to reassure students. They could promise to collect only data that has a tangible connection to students’ academic and social success. They can pledge not to sell or share the data with third parties. And they can offer to destroy data that is no longer valuable, such as dining-hall entry information for students who have already graduated.

Facebook wants its users to share more about their personal lives

Formalizing these standards in a code of ethics could make students comfortable with increased data collection and make colleges accountable for the appropriate use of these data.

Private companies and nonprofit organizations already employ big data to drive sales and predict consumer behavior.

It’s about time colleges apply these same technologies to help students stay in school and reach their full potential.