INDIANAPOLIS — A trio of senior college enrollment officials gave a peek into how they decide which students to recruit. The process now involves number-crunching students’ demographic and economic information — not just sending chipper ambassadors to every nearby high school, mailing glossy books to students’ homes and relying on gut instincts.
The discussion, during a session at the annual meeting of the National Association for College Admission Counseling, was one of many to take place here about how to hunt for students. The search for students involves a web of data points, formulas and consulting firms that perhaps few parents and students are aware of.
Don Munce, the president of the National Research Center for College and University Admissions, or NRCCUA, offers a modeling service meant to predict which high school students are most likely to enroll at a particular institution. The center sells data on students to college admissions officials.
Munce moderated the panel of three college admissions officials who use his predictive modeling service. One of the college officials joked he bought so many student names from NRCCUA that he probably paid for Munce’s yacht.
Munce advocates a “smart approach” — which is the brand name of the modeling service he sells — that would help colleges target the students most likely to enroll.
Are officials targeting students likeliest to attend their college, “Or are you throwing a lot of activity out there at students who don’t enroll or may never enroll?” Munce asked.
NRCCUA, the College Board and ACT all gather data on students. The three services sell millions of student names for about 37 cents apiece to colleges and consulting firms hired by the colleges. They also all sell colleges predictive modeling services based on the data they collect — things like family income, GPA and zip codes. NRCCUA gathers data on students through surveys they are asked to take in high school and has about 300 data points it can use.
There’s a funnel for prospective students that admissions officials — and chief financial officers and presidents — pay a lot of attention to. It starts with a general population of students. Then there are inquiries, which means the students who ask a college for information. Then there are those that really apply, those that are actually admitted and those that finally attend.
The funnel can be huge at the top. High Point University, a private college in North Carolina, gets inquiries from 70,000 students. This fall, 1,370 freshmen enrolled, said Andrew Bills, High Point’s vice president for enrollment.
High Point has grown rapidly. In 2005, its freshmen class was just 412.
Bills said the university decides what prospective students to focus on using modeling.
About 90 percent of the students who enrolled at High Point came from the 40 percent of students a model scored above a .66 on its scale going to 1. The factors include geography, academic interest, whether students would prefer a private college and family income. So, for the university, it makes sense to focus its recruitment effort on those students, rather than students with a lower score.
“We were wasting our time messing with a lot of those inquiries,” Bills said.
Stephen Lee, the executive director of admissions and recruitment at West Virginia University, uses modeling to help sort through the pool of out-of-state students and find out which out-of-state students recruiters should spend time communicating with.
“We wouldn’t be able to do that if our mission was, ‘Get to every high school,’” Lee said. He showed a chart that had some Pennsylvania students — a key feeder state for West Virginia. And that showed how many prospective students went to each high school and of those students how many had a predictive modeling score that showed they were likely to attend.
James Steen, the vice president for enrollment management at Houston Baptist University, uses modeling to target students but also to figure out what traditional recruitment methods work or not.
Houston Baptist used to send out about 12,000 viewbooks — those glossy guides aimed at prospective students that cost a lot to make and mail. One year, Steen decided to not send viewbooks to 2,000 of the students his model predicted were most likely to attend.
He found would-be students in the group that didn’t get the viewbook actually were more likely to attend than those that got one.
“So guess what? We didn’t do a viewbook last year and we lived to tell about it — our enrollment was still up this year,” Steen said.
The ability to slice and dice prospective students has helped some colleges diversify their class and target students who might not think about going to their college or even college at all, but also raises the specter that some colleges with worrisome bottom lines will only go after certain kinds of students — like wealthy students who can pay their way without scholarships — at the expense of others. Traditional recruiting already targets wealthier families in some cases. For instance, the Los Angeles Times found that college recruiters gave low-income high schools fewer visits than affluent high schools.
Advisers to the College Board -- which has data on seven million students it sells to about 1,100 institutions each year – met last summer and talked about doing more to police how colleges can use the board’s student data, but a committee decided not to change its policies.
Want articles like this sent straight to your inbox?Subscribe to a Newsletter