Big Data in College Recruiting

A Stacks of books and a graduate cap

Lists of prospective students have become fundamental in college recruiting. University of Arizona research has exposed how the data and tools that produce these lists routinely marginalize already underrepresented populations.

Karina Salazar, assistant professor in the College of Education, partnering with Ozan Jaquette at UCLA, obtained the details of 830 list orders from 14 public universities, including which filters vendors used to match students to those orders. They also analyzed the resulting roster of more than 3.6 million then-prospective students.

While methodologies vary slightly from one list vendor to the next, all offer ways to sort students not simply on grades or test scores but also on factors such as how many other students from their high school or zip code historically attended college – criteria highly correlated with race and household income.

The studies found that these filters dramatically reduced diversity in resulting lists. They also found that associated travel for recruiting overwhelmingly focused on higher-income, mostly white communities.

Salazar notes that university staff often outsource list purchases. They may have little understanding of how hired consultants and list vendors translate their expressed goals into the lists they get back.

The underlying complication, she argues, is that structural racism persists in college recruiting precisely because lists and filters viewed as “normal” systematically benefit populations that already dominate higher education.


Data Connects Us Magazine