Predicting Student Success
Southern's academic advisers typically provide students with the benefit of their knowledge and experience when guiding them into and through a degree program. But soon, they will be able to "look into the future" and offer them statistical-based evidence as to whether they will likely succeed in a specific major.
The university will use a predictive analytics system, developed by the Education Advisory Board in Washington, to help students weigh their options when considering a degree program. The system uses data, particularly student grades, to predict whether someone is likely to succeed in various majors and courses. The university is using the model as part of an effort to boost its graduation rate, as well as the various retention rates.
"This model has a good track record in terms of projecting whether a student will do well or have a difficult time in a specific class or degree program," says Richard Riccardi, director of Southern's Office of Management Information and Research.
"That's not to say the results will be perfect, or that a student cannot make up his or her own mind whether to pursue a particular degree program. But the analysis has proven to be accurate a very high percentage of the time."
As an example, if a student has gotten C's in his or her math classes, the chances for success as a business management major are slim, Riccardi says. "The new program will enable an adviser – who might already be skeptical that a student would succeed in a business management program – to point out the hard data to the student, as well as offer some alternative majors in which they are likely to be more successful.
The hope is that students will learn at an earlier stage in their academic careers whether a particular program is a good choice for them, Riccardi says: " If it is not, switching to another major could very well make the difference between completing a program and graduating, or not. Or, it can make the difference in whether someone graduates within six years."
Kimberly Crone, associate vice president for academic student affairs, says she is optimistic that the tool can begin to be used for the 2014-15 academic year. "We have to submit 10 years of data before the system will be operational," she says. "But once that information is plugged in, this program will be a real benefit to our students."
The system also can be used to determine the level of difficulty a student is likely to have in a course. Riccardi notes that sometimes an adviser might not be aware that poor grades in some courses predict a lack of success in an entirely different area. He said this program takes much of the guess work out of the equation.
"In some ways, this system is like 'Netflix on steroids,'" Riccardi says. "After data is obtained, Netflix can pretty accurately predict whether you might like a particular movie based on what you have watched previously. It's somewhat similar with this program in terms of predicting success in courses and majors."
Riccardi adds that the "human element" remains important, noting that there might be a legitimate reason why a student didn't do well in certain courses that doesn't reflect on their aptitude.