How Ole Miss Is Using Data to Identify Students at Academic Risk

by | Jan 23, 2019 | Education

Reading Time: 4 minutes

Recently EDUCAUSE spotlighted institutions where IT leaders have been helping their organizations make increasingly information-based decisions. The University of Mississippi was featured as one case study in how institutions of higher learning can use data and analytics to improve student outcomes. The university has 23,000 students across all of its campuses, and with its 55 percent overall increase in enrollment over a ten-year period ending in 2016, it is among the fastest-growing colleges in the country.

That rapid growth, combined with UM’s diverse student body and relatively limited resources as a public university in a poor state, means that the university needs to take advantage of every opportunity it can to get ahead – and data can help it do that.

Laying the groundwork

The University of Mississippi has been laying the groundwork for years to prepare for the type of information it is now getting from its data. It has gradually been able to build upon the infrastructure it had in place to gather different kinds of data, such as student attendance information collected from automated scanners in classrooms that report absences to the system in near-real time.

Over the years UM has created data analyst positions to help go through the data, and it has increased data science training and certifications among the staff.

Using the data

Many schools struggle with what to do with the information once they have it. There can be an overwhelming amount of data and it can be hard to zero in on areas where the data can be practical. Ole Miss, with the foundation established by its IT department, had no such trouble, using the data to focus on student success.

As a starting point the school looked at its classification of ‘academic risk.’ The school has a strong advisor program and those advisors can look at how students are doing through the school’s on-line portal. Absences, instructor alerts, and midterm grades were the factors that went into letting an advisor know if a student was in the academic risk category – at risk for not completing a fall or spring term with good academic standing. Many advisors, though, thought the midterm grades came too late to intervene and help the student.

Solving a problem

The university recently changed the process to use predictive analytics. UM worked closely with the school’s Center for Student Success and First Year Experience to increase the variables that would indicate an at-risk student to ten – in addition to absences, instructor alerts, and midterm grades they included elements like high school performance in core subjects for new students and most recent GPA for continuing students – to predict the probability of a student having “good” or “not good” academic standing. They worked to increase the accuracy rates of their data model, and they created an academic risk dashboard for advisors.

The filtering process is entirely automated, saving time for the IT department in what would be an otherwise arduous manual task.

The results have been positive so far, but like any good work with data, the university is already thinking about how it could be better. Since the results of the work so far mean there are actions taken for students if they are considered to be at academic risk, advisors suggest collecting data on those interventions. By measuring the types of intervention and their effectiveness, the University of Mississippi can use data further to increase the impact data can have on its students’ lives.

Related articles

John Sucich
Follow me

You may also like