Meet the Secret Algorithm That's Keeping Students Out of College

In March 2020, the International Baccalaureate (IB) organization canceled its spring exams due to COVID-19 and deployed a statistical formula developed by an unnamed educational data analysis organization to calculate final grades for over 170,000 students. The system used student assignment scores, predicted grades, and historical IB results from their schools to predict what students would have scored on the canceled exams. The algorithm created 'a bespoke equation' for every school, and for schools lacking historical data, predictions were built using pooled data from other schools. When results were released on Monday, many students received suspiciously low scores that were much lower than their teachers' predictions. Students like Anahita Nagpal in Germany lost university places and scholarships, with Nagpal's acceptance to NYU's premed program being jeopardized. More than 15,000 parents, students, and teachers signed a petition asking IB to revise their grading algorithm. Teachers reported that nearly all their students were downgraded compared to predicted grades, with some students losing scholarships to top universities. The IB foundation declined to provide details about the methodology but said the system had been checked against five years of past results and that disappointed students could use the existing appeals process for a fee.