A flood of students unexpectedly accepted admission offers. A UC campus was caught off guard. Administrators scoured the files of the admitted and took a hard line on those who had failed to meet paperwork deadlines. They withdrew more than 500 offers, causing a furor.
The year was 2015, the campus Santa Cruz.
The storm that UC Irvine recently unleashed when it took a similar approach to overenrollment was unusual but hardly unheard of on the nation’s college campuses. Experts say the two UC cases and others like them at Temple University in Philadelphia and St. Mary’s College of Maryland underscore the vagaries of enrollment prediction — a discipline that aims to meld the science of data analysis with the guesswork of anticipating teenage whims.
Cal Poly San Luis Obispo also miscalculated its numbers this year, with about 700 more students saying yes to its offers than expected. One curveball was a campus decision to eliminate the option for students to make an early, binding commitment to enroll, which boosts their chances of admission but was seen as advantageous to the wealthier ones who did not need to wait for financial aid packages.
The move made a difference, with a 30% increase in the incoming freshman class of low-income and underrepresented minority students. But it also made the process more unpredictable, since early commitments previously had locked in nearly a quarter of the enrollment early on, said James Maraviglia, the associate vice provost who oversees the admissions process.
It wasn’t just freshmen, who accepted admission offers at a far higher rate — 34.6% this year compared to 29.5% last year. Four hundred more current students than anticipated said they were planning to return to campus, many of whom benefited from enhanced academic support and counseling to keep from dropping out, said Mary Pedersen, senior vice provost of academic planning and programs.
Unlike Irvine, however, the popular California State University campus did not face the crunch by trying to find ways to withdraw more admission offers. Instead, it is working to accommodate the extra students by reconfiguring dorms to add 1,000 more beds and hiring additional faculty.
“Enrollment management is an art, not an actuarial science,” Maraviglia said. “One variable can change and totally change your mix.”
He and others said predictions have become even trickier since the advent of online applications has made it easier for students to apply to multiple campuses.
Tom Green, a college enrollment management expert who has consulted with several Cal State campuses, said California predictions are made particularly difficult by the sheer size of its three higher education systems, which collectively educate 2.9 million students.
“How do you try to predict things ...when you are so much larger than anybody else that there’s no other model like you?” said Green, associate executive director of the American Assn. of Collegiate Registrars and Admissions Officers. “Trying to manage enrollment and resources at that scale is really challenging.”
UC Irvine, in the end, has had to reverse course. Stung by public outrage over its abrupt cancellation of nearly 500 admission offers after 850 more students than expected said they would attend, the school announced last week that it would readmit most of those affected. Chancellor Howard Gillman apologized and announced that he had directed the internal auditor to review the admissions process to try to understand how officials so badly underestimated the enrollment, why students who got the notices were poorly treated when they inquired about them and how to avoid repeating the mistakes.
UC Santa Cruz launched a similar review after its 2015 fiasco, which prompted some outraged parents to contact their state legislators to demand answers and changes in procedure. The campus subsequently overhauled not only how it predicts its enrollment numbers, but also how it publicly explains its admission process and handles those who fail to meet paperwork deadlines, said Michelle Whittingham, associate vice chancellor of enrollment management.
Those changes seem to have made a difference, she said. UC Santa Cruz missed its mark by 366 students in 2015 — but by only one in 2016 and seven in 2017. And thanks to more forgiving policies adopted last year, campus officials this year revoked only 31 admission offers for paperwork problems, such as missing deadlines for submitting transcripts.
“We recognize this was exceptionally devastating to students and families,” Whittingham said of the 2015 cancellations. “We really tried to dissect every piece of our process and do everything we can do to make sure nothing like that ever happens again.”
For starters, the campus zeroed in on weaknesses in its forecasting model. Most campuses start with enrollment targets and work backwards, calculating how many students need to be admitted to reach them by looking at such factors as previous rates of acceptance, known as yield, and of students who initially accept but then decide to enroll elsewhere.
At Santa Cruz, officials used to predict group behavior. Now their model looks at each applicant, calculating the likelihood that a student will come based on such variables as their grades and test scores, zip codes, intended majors and campus visits.
Those considered more likely to attend UC Santa Cruz generally live closer to campus, visit more frequently before deciding and have a weighted grade point average that is somewhere around 3.9. Students planning to major in engineering, film and computational media are more likely to attend than those focused on music or feminist studies.
Not that the carefully crafted forecasting model can account for all teen decisions. When Whittingham asked students why they chose UC Santa Cruz, one told her “oxygen levels,” and another said the campus was close to Seattle.
“One of the things you can’t lose sight of is that literally we’re predicting the behavior of 17- and 18-year-olds,” she said. “I can’t predict my own son’s behavior.”
The campus model has a built-in margin of error that admits fewer students and places more on a waitlist. That way, overenrollment is less likely and wait-listed students fill the gap in cases of underenrollment, Whittingham said.
Other campuses also rely heavily on waitlists, though some use them differently. Cal State Fullerton, for instance, offers those who miss paperwork deadlines spots on the waitlist rather than completely pulling their offers. The campus needs to enforce deadlines because it’s so popular and, for years, has had to turn away thousands of eligible students, said Darren Bush, interim associate vice president of student affairs. With the help of regular data updates throughout the admissions process, officials have fine-tuned their forecasting model to be on target this year.
More information sharing between the UC and Cal State systems would be helpful, Whittingham said, since all of the schools are affected by the actions of their competitors. Santa Cruz’s lower-than-expected yield rate this year might have been affected by Cal Poly SLO’s elimination of the early decision option and subsequent enrollment surge, she said.
Since 2015, officials from UC campuses have gathered at an annual summit to share their enrollment targets and forecasting strategies. That’s helped with planning, Whittingham said. When UCLA and UC Berkeley announced they would increase admission offers last year as part of the UC system’s pledge to add 5,000 California freshmen, she knew fewer applicants would accept offers from her less-selective campus.
To improve its process, Santa Cruz also has stepped up its outreach to students, parents and high school counselors, sending out more reminders about paperwork deadlines and being less severe about them. The campus extends the July 1 deadline for transcripts to Aug. 10 for those who need more time, for instance, and has restarted mailing reminders to parents after many said they never saw the electronic information sent to students’ online portals and email accounts.
That said, no amount of planning is fail-safe.
“People spend years trying to create predictability out of unpredictability,” Whittingham said. “So much of this work is illogical.”