The Precarious Nature of College Admissions Algorithms

The college landscape has dramatically shifted over the past half century with effects that can clearly be seen. Tuition rates are up, and colleges are in extreme competition with one another for ever precious enrollment. Recently, a new weapon has entered the fight between colleges for prospective students. Enrollment algorithms are being used at hundreds of institutions. But as a paper from Alex Engler recently noted, there are many potential problems with their application. Let’s take a look. 

What Is Enrollment Management? 

Before diving into college admissions algorithms, we need to first visit the topic of enrollment management, the fancy professional term for college admissions. Since the 1960s, government funding for colleges and universities has seen dramatic decreases. As a result, universities see a larger and larger proportion of their bills paid by students through their tuition payments. 

Thus, each year colleges carefully allocate expected funding with a prediction of how many incoming students they will get. This involves a number of factors such as yield rate; however, the bottom line is colleges have immense pressure to deliver the number of students predicted. Falling short of that would mean the college potentially operates in the red. 

What Are Admissions Algorithms? 

You likely have heard the term algorithm and suppose it has something to do with math and computing. You’d be right! Algorithms are often complex mathematical formulas and processes that are run by computers in order to help with decision making. 

In fact, algorithms are a part of your daily life. Your e-mail spam filter? Likely powered by an algorithm that classifies incoming e-mails as spam or not. Ever receiving an automated call or text from your credit card company about suspected fraud? An algorithm at work. Algorithms are used in many ways. 

Admissions algorithms seek to help colleges optimize their recruitment of students by doing two things. First, they are predictive in that they attempt to use vast arrays of past admissions data to predict the likelihood that a student will accept an offer of admission. Secondly, they create optimization. This means that they adjust the amount of financial aid to offer students based on how much it is predicted to take for them to accept an offer. 

What Are Potential Problems with Admissions Algorithms? 

There are a number of potential problems that may arise – and already are arising – from the use of these admissions algorithms. Let’s take a look at some of the most pressing concerns that have been expressed in the paper by Engler as well as by other experts in the field. 

Too Narrow of a Focus. Many people have critiqued admissions algorithms for being too narrowly focused. They aim to determine whether or not a student will accept an offer of admission and attend the university. However, they do not consider whether or not a student is predicted to be ultimately successful at that university. 

This creates a potential risk of admitting students who are not academically prepared and do not have the support necessary to succeed. Part of this is because the business side of universities (i.e. enrollment management and financial offices) and the success side of universities (i.e. student life and faculty departments) operate as silos with little communication. 

However, the reality remains that most universities are experiencing problems with retention and graduation rates. Admitting students who will ultimately not graduate is simply shifting a greater burden to the student. 

Increasing Existing Barriers. Much of higher education’s focus during the past few decades has been to reduce existing barriers in access to higher education. The field has recognized for some time that there are significant barriers faced by students of color, students from lower socioeconomic status, and first-generation college students. 

Most colleges have created a number of programs aimed at attracting and preparing these students for success once on campus. However, the use of college admissions algorithms may increase gaps faced by these students. An algorithm is very efficient at its purpose. If it finds that Latino students or poor students are less likely than white students or wealthy students to accept admissions offers, it may respond by ultimately deprioritizing recruitment of those students. 

Increased Student Debt. One of the most pressing concerns voiced regarding admissions algorithms relates not to their predictive function but rather their optimization function – the part that adjusts the amount of financial aid offered to increase the likelihood of acceptance.

For some students, this means that the algorithm will offer more aid. However, for many students, this means the algorithm will reduce the amount of financial aid they would have been offered. This is because the algorithms have a double goal: increase acceptance rates while using the least possible amount of funds. 

If the algorithms are truly successful in this, it means many students will receive less aid and incur more debt. This will only serve to amplify the existing crisis of student indebtedness, a concern that everyone should be paying attention to. 

What Does This Mean for Students? 

The reality is that college admissions algorithms are likely here to stay. As noted, there is strong competition between colleges while reduced governmental support means tuition is critical for meeting operating budgets. 

The best-case scenario is that there will be careful attention paid to potential unintended effects and algorithms will be adapted to reduce these effects. However, the reality is that some of these potential problems will come to light. Will colleges act to address them? That will remain to be seen. Shamrck can help you figure out if college is even for you. And give you the resources to better stand out to these algorithms. 

Visit the Shamrck Dashboard today to let us help you with your journey to higher education.