Handcrafted: Enrollment Managers Use Predictive Modeling Tool to Create a Class

ACT’s predictive modeling tool offers statistical analysis of information students supply when they take the ACT Assessment—demographics, enrollment preferences, and academic ability. Postsecondary institutions can use data from the students who have already expressed an interest in the school—the "score-senders" model—or data from all ACT-tested students through ACT’s Educational Opportunity Service. In either case, because they are based on student-level information, predictive modeling data are more accurate and less expensive than predictions based on traditional demographic data based on geography.

“Student recruiting is so expensive per student that institutions really want to fine-tune it,” said Michael Hovland, senior consultant at ACT.

Blackburn-Smith agrees, “We all are always looking for ways to be more efficient.” Ohio State has used pilot versions of ACT’s predictive modeling tool for several years, as ACT refined it. Blackburn-Smith said it took awhile for the staff at his university to be comfortable using it.

“There comes a time when an administrator just has to trust the model. The first time you say ‘I’m going to mail to 30,000 fewer students than I usually do,’ it’s very scary,” he said. “We found that we could be very targeted using predictive modeling.”

The university’s ability to target the best students has helped it raise its average ACT Composite score for entering freshmen from 22.8 to 25.4.

“We are very pleased with what we’ve been able to do with predictive modeling. For the last eight years, each incoming class has been the best ever,” he said. “We’ve increased diversity while improving the academic profile, and we’ve maintained socioeconomic diversity as well.”

Blackburn-Smith says his university will continue to use ACT’s predictive modeling tool. Now that it is available to all postsecondary programs, he encourages other institutions to take advantage of it, too.

“It really is a tool that will give you a competitive edge over folks who are not using it, and more and more are using it, so you’re at a competitive disadvantage if you’re not,” he said.

For more information on ACT’s Predictive Modeling for Recruitment and Retention, contact Michael Hovland at 319/341-2295.