Test-Optional: Who Needs Test Scores?

Test-optional admissions is not a new discussion in higher ed. According to FairTest, the list of colleges going test-optional for admissions is growing. Higher ed institutions have cited several reasons why they do and don’t use test scores for admissions and merit aid decisions. In 2020, the COVID-19 pandemic forced the hand of many colleges who used test scores to move to test-optional for the short term or permanently. High school students simply couldn’t and, in some cases, still can’t take the tests.

No matter the reason, higher ed institutions need to assess and refine, as needed, their procedures in evaluating applications and managing merit aid in a test-optional world.

How One Institution Approached Test-Optional Admissions for the
Class of 2021


A higher ed institution (and Othot partner) has historically required test scores in combination with other academic profile indicators to evaluate a student for admissions. Students who met specific thresholds were admitted. Now, the school is not requiring test scores for admission.

We’ve been working with this partner institution and others to tackle this dilemma quantitatively, and we looked at several approaches, including:

  1. Replace test scores with other variables.
  2. Develop a prediction for each student that answers the question: “What is the likelihood of this student to be admitted from the applicant phase?”
  3. Predict a test score. *

In this situation, the institution implemented the second approach and a more detailed application reading process.

For this higher ed institution, we created a prediction for each applicant that answers the question, “What is the likelihood of this student to be admitted from the applicant phase?” Using historical data (and excluding test scores), we modeled all applicants who were admitted and who were not admitted. Then, we built a model to predict whether or not the students currently in the applicant pool would be admitted.

The admission prediction became an additional data point for the college to use when evaluating a student for admission.

*Our data scientists tested approach #3, “Predict a test score.” We do not recommend it as a primary approach because of concerns about bias in the tests. If you’d like to learn more about predicting test scores, please watch our webinar, “The 5 Questions to Ask if Test-Optional is Best for Your Institution.”

Test and validate the models

We tested and validated the predictions to make sure that the institution understood the risks of this approach. This is a critical and important step as higher ed institutions think about ways to review applicants without test scores.

Without testing and validating the models, a university could admit too many students or shape the class in a way that does not align with goals, which only emphasizes the need to test and validate the models to determine the best way to apply these admission predictions.

But what about merit aid awards and test-optional admissions?

How One Institution Approached Merit Aid without Test Scores for the
Class of 2021


Many higher ed institutions have a defined financial aid matrix that relies on test scores, class rank, GPA, and/or other academic indicators to determine how best to leverage merit aid to meet institutional goals.

With a set merit aid budget or net tuition revenue goal and defined enrollment goals related to factors such as quality, residency, and diversity, how can the institution award merit aid without a test score?

Like test-optional admissions, we looked at a few approaches with one of our partner institutions:

  1. Replace test scores with another variable.
  2. Derive an additional variable to use, such as a predicted test score, academic rank, or even predict what a historic award may have been from a previous strategy.

In this example, the college implemented the first approach.

We worked with our partner institution to identify another academic profile variable that could replace the test score in the merit aid matrix. The merit aid matrix was revised with the new variable. We applied a similar population distribution to the new matrix based on the population distribution from the strategy that had included test scores.

Test and validate the models

Then, with the college, we tested the models to identify any risks in this new approach, e.g., overspending their financial budget or yielding a class that doesn’t align with their overall enrollment goals. The validation step is critical because, with any new approach, it is important to understand the risks and any steps that need to be taken to mitigate those risks.

What’s Your Approach for Test-optional?

FairTest reports that two-thirds of all U.S. colleges and universities are going test-optional for the class of 2021, and the change has major ramifications for higher ed, students, and their families. This change may be short-term or long-term and necessitates that institutions evaluate their actions and processes. In our experience with our partner institutions, we’ve learned that there is not a one-size-fits-all solution.

The higher ed institution must have clarity related to enrollment goals. Then, the institution must decide upon a course of action, analyze that course of action, and then identify and mitigate risks. That's what we are doing to support our partner institutions, and we can support your institution too.

If you’d like to learn more about our approaches to test-optional admissions and merit aid, please contact us at othotteam@othot.com.