Inaccurate, impossible: Expert knocks on the door New Trump plans to collect university admissions data

“You’ve worked hard to add these elements without having to review the mechanisms of new variables, and the system that ensures they are properly implemented,” Cook said. “You’d almost think that the people who are doing this don’t know what they’re doing.”

Cook has helped provide education departments with IPEDS data collection for 20 years and serves in the technical review team, often convening the data first to suggest changes to the data collection. The groups were disbanded earlier this year, without a set up of a new admissions data proposal for veterinarian Trump.

Cook and other data experts can’t figure out how the deprived educational statistics agency handles the task. All six NCE employees involved in the IPEDS data collection were fired in March, while only 100 employees in NCES run by one agent specialist also had other jobs.

An Education official who did not want to be named denied that no one left behind the education department’s experience. The official said staff in the Chief Data Office, separate from the Bureau of Statistics, were “very familiar with IPEDS data, collection and use.” Former education staff told me that some of them had experience analyzing data but did not collect the data.

In the past, as many as a dozen employees have worked closely with RTI International, the Institute of Science Research, which handles most of the IPEDS data collection efforts.

Eliminate technical reviews

Particular concern is that RTI’s $10 million annual contract was cut by about half by the government’s Department of Efficiency (also known as Doge), two former employees who demanded to remain anonymous out of fear of retaliation. Those serious budget cuts eliminated the technical review panel for reviewing IPED proposed changes and ended training on the correct submission of data by colleges and universities, which contributed to data quality. RTI did not respond to my request to confirm the cuts or answer questions about the challenges it faces in terms of budget reduction and staffing.

The education department has not denied that the IPEDS budget has been cut in half. “The RTI contract focuses on the most critical task IPEDS activities,” said the Ministry of Education official. “The contract continues to include at least one task that can convene a technical review panel.”

Other elements of IPEDS data collection have also been reduced, including contracts that check the quality of the data.

Last week, the scope of the new mission became even more obvious. On August 13, the government released more details on the new admission data it wanted, describing how the education sector was trying to add a brand new survey to the IPEDS called the Supplementary Admissions and Consumer Transparency (ACTS), which would categorize all admission data, as well as the outcomes of most students, the outcomes of most students, and the financial aid data of race and members. The university will have to report on undergraduate and graduate school admissions. The public has 60 days of comments and the government hopes universities will start reporting this data this fall.

Complex collection

Christine Keller, executive director of the Institutional Research Association, which collects and analyzes data, calls the new survey “one of the most complex IPEDS collections ever.”

Traditionally, it has taken years to make smaller changes to IPED and given universities a year to start collecting new data before submitting it. (About 6,000 colleges, universities and vocational schools must submit data to IPED as a condition for students to obtain federal student loans or receive federal Pell grants. Failure to comply with the threat of fines and loss of federal student aid.)

Typically, the education department reveals screenshots of the data fields to show the university required to enter the IPEDS computer system. But the department has not done so, with some of the data descriptions being ambiguous. For example, colleges will have to report one in five test scores and GPA and be broken down by race, race and gender. One explanation is that a university must say, for example, how many black male applicants scored in the 80th percentile on the SAT or ACT. Another explanation is that universities need to report the average SAT or ACT scores for the top 20% of black male applicants.

The Institutional Research Association once trained university administrators how to properly collect and submit data and categorize it through confusing details – until Doge eliminated that training. “The lack of comprehensive, federally funded training only increases the risk of institutional burden and data quality,” Keller said. Keller’s organization is now offering a small amount of free IPEDS training to universities.

The education department also requires universities to report five years of historical admissions data, divided into many subcategories. Institutions have never been required to retain data from unregistered applicants.

“They were asking for five years of previous data, which was incredible,” said Jordan Matsudaira, an economist at American universities. “It would be a square in the pandemic era when no one reported the test scores.”

“Misleading results”

Matsudaira explained that IPEDS had considered asking universities to solicit more academic data on past racial and racial data, and the education department eventually rejected the proposal. One problem is that slicing data into thin slices and breaking it down into smaller and smaller buckets will mean that there will be too few students and that data must be suppressed to protect student privacy. For example, if there are two Native American men in a college’s SAT score, many people might guess who they are. A large amount of suppression of data will reduce the usefulness of the entire set.

Moreover, a small number of numbers can lead to weird results. For example, a small university can only have two Hispanic male applicants, with a high SAT score. If both are accepted, that is 100% enrollment. If you accept 400 white women with the same test scores, there is only a 50% admission rate. On the surface, it looks like race and sexism. But this could be a fluorine. Perhaps these two Hispanic men are athletes and musicians. The next year, the school may reject two different Hispanic male applicants, but not so impressive as extracurricular. The admission rate for Hispanic men with high measured scores will drop to zero. “You end up having misleading results,” Matsudaira said.

Reporting average test scores by race is another big concern. “To me, it felt like a trap,” Matsudaira said. “Mechanically, this will make the government pretend to claim that black students have lower admission standards relative to white students when you know that this is not the right inference at all.”

The statistical problem is that at the high end of SAT score allocation, there are more Asian and white students, all of these perfect 1600s will provide averages for these racial groups. (Like a very tall person, the average height of a group of people will be biased.) Even if a university’s test score threshold applies to all racial groups and no one admits that people under 1,400 are still lower than those of white students. (See the graph below.) The only way to avoid this is to accept it purely through the test scores, and only the students who get the highest scores. In some highly selective universities, there are enough applicants with 1600 SAT to fill the entire course. However, no institution can fill its student body by simply passing the test scores. This could mean ignoring applicants who are likely to become concert pianist, celebrity footballer or outstanding writer.

Average Score Trap

Admission data is a highly charged political issue. The Biden administration initially led university admissions data by race and race. Democrats want to collect this data to show how universities and universities in the country have become more and more as affirmative action ends. After a complete technical and procedural review, the data is scheduled to begin this fall.

Now, the Trump administration’s requirements have been done and many new data requirements have been added without following the normal process. Rather than tracking the diversity of higher education, Trump hopes to use admissions data to threaten universities and universities. If the new instruction produces unfamiliar data that is easily misunderstood, he may get his wish.