UMN works to address questions about admissions process

In an audit released last year, fifteen recommendations were made to resolve “essential” issues in the admissions process.

Helen Sabrowsky

Following an audit that outlined fifteen areas of improvement last year, the University of Minnesota’s Office of Admissions is still working to resolve four “essential” recommendations, including improvement of accuracy and consistency of data used when making admission decisions. 

The office has addressed 11 of the initial 15 recommendations so far, through implementing documentation of final admission decisions and improving the review process to minimize the subjectivity of decisions along with other changes. 

The audit team reviewed 58 student applications to the University from fall 2016 that fell into six categories. One category included 10 admitted applicants with low test scores, GPAs and/or class rank. The decision to admit these students were below standards of typical admitted students. The audit found that two of these 10 students seemingly should not have been admitted, and the reasoning for admission was not documented.

“You can’t really tell how they’re admitting students and whether they’re being fair,” said Regent Michael Hsu.

Hsu criticized the audit’s small sample size and encouraged the team to follow up on the initial study using a larger number of applications so they could better determine the scope of the issue. He said the Office of Admissions moved forward in addressing the recommendations without additional investigation. 

Because the University was inconsistent in documenting why students were admitted or rejected, Hsu said it is impossible to determine whether the University is being unbiased in admissions.

This stems from the subjective nature of the University’s holistic application process, said Heidi Meyer, executive director of the Office of Admissions.

The process admits students based on primary review factors like test scores and grade point average, as well as secondary characteristics such as leadership experience and military service, she said.

“We may make some admission decisions that look out of variance, but in the context of that student’s application file, we’re looking for factors of academic success,” Meyer said. 

The audit identified four secondary factor codes as being open to interpretation, and the Office of Admissions expects reviewers to use their best judgment for these codes. The audit found “several instances” where the codes were interpreted differently among reviewers, which the audit said could lead to inconsistency in the review of applications. 

The University has refused to say how much of an impact secondary factors have during the admissions process or whether they can override a deficiency in primary factors, Hsu said. 

Regent Linda Cohen, chair of the Board of Regents audit committee that Hsu also sits on, acknowledged that the holistic admission process can be subjective.

“When one does this holistic review, it’s not as straightforward or simple as ‘somebody met a [certain] ACT score’ … it’s in the context of other things, and I believe in that process,” she said. “Is it perfect? No, because there are some judgment calls being made there.“

She stressed the importance of implementing more consistent documentation.

“To me it wasn’t the process as much as how you document those decisions,” she said. 

The Office of Admissions, which was aware of the issue at the time of the audit, has now resolved the audit’s recommendation to improve documentation. 

Hsu said the University prides itself on meeting numbers and filling colleges, but they don’t keep track of how they meet enrollment goals. He said the University was inconsistent in converting applicants’ GPAs to a standard four-point scale, which he said he found concerning. 

The audit states that 76 manually-entered data points, such as GPAs, secondary factors and high school rank were entered incorrectly by first level readers. 

“It is imperative to have accurate and consistent data in all locations when critical data is stored in multiple systems. If inaccurate information is stored in one or all sources, this could lead to issues with data integrity and could contribute to future inaccurate admission decisions,” read a statement from the Office of Internal Audit in a December 2017 Regents docket.

This recommendation has not been resolved by the Office of Admissions, though it is in the process of implementing an automated interface between all systems used in the admission process. 

“I think all applicants should be treated fairly, and it’s not clear to me that they’re being treated fairly,” Hsu said.