Using Big Data to Improve First Year Student Retention

Feb. 7, 2017

Sudha Ram, Ph.D., the Anheuser-Busch Chair in MIS, Entrepreneurship and Innovation, was recently quoted in the New York Times discussing her research using big data to improve first year student retention rates at universities.

"About 20 percent of first year students drop out at the end of each year – and not necessarily because of grades – but it's very hard to find out who is going to drop out until they do," said Dr. Ram, who is an Eller MIS professor and the Director of the INSITE Center for Business Intelligence and Analytics.


"It turns out that students who are dropping out don't wait for their grades," she said. "They decide that they are going to leave 10-12 weeks into the first semester, and it's very hard to figure out who these students are because we don't have their grades yet. So how do we figure it out? You can’t ask them, because the ones who drop aren't going to respond, so we said let’s use big data."

As a big data researcher, Dr. Ram mines for patterns in very large data sets to connect different types of data together to provide insights.

"The general principle we follow is to identify a question or a problem and then figure what sorts of data sets we can use and connect to produce signals that would answer that question," she said. "And it's not just to look at the past; with big data, now you can do predictive analytics and make predictions about the future." 

In this study, Dr. Ram relied on anonymized Smartcard data to get signals about how students were integrating into campus life. Students use their Smartcard to access labs, classrooms, libraries, services in the student union, vending machines, and more, so Dr. Ram was able to extract movement and behavior data based on Smartcard information and get signals about levels of integration. 

"We built a model to predict which individual first year is at risk of dropping out and we can apply it about 12 weeks into the first semester," she said. "We take their Smartcard information, we look at changes in their social interaction behavior, we look at how well integrated they are into campus by what services they use, and then we use demographics, but we don’t use any academic indicators."

The results can have a major impact on future first year student retention rates.

"We are able to predict 90 percent of all of the students who drop out – the actual individual students," she said. "If you can predict which student is likely to drop out, then you can do interventions to help him or her not drop out." 

"We are proud of Dr. Ram's research as McGuire Center's Anheuser-Busch Chair in MIS, Entrepreneurship and Innovation," said McGuire Center Director Remy Arteaga. "This research takes an innovative approach to tackling the difficult issue of student retention, and the results could aid educators in finding ways to help students engage and remain in school." 

This goal for research to have social impact is a common thread among many of Dr. Ram's projects. She has received attention most recently for research that uses big data to address transportation troubles in Brazil's fifth-largest city, and to predict asthma-related emergency room visits.

"Everything I do is about data and figuring out insights from it," she said. "Big data analytics is really about innovative thinking and different kinds of ideas. It's basically solving a problem in a new and different way. I try to look at a problem and come up with new, innovative solutions, and I like to do research that has some kind of social impact." 

Read the New York Times article: Will You Graduate? Ask Big Data