Aspiring students who wrote content-rich university admission essays were more likely to end up with higher grades in their classes1.
Jonah Berger at the University of Pennsylvania in Philadelphia and Olivier Toubia at Columbia University in New York City used machine-learning techniques to analyse more than 40,000 essays written by applicants to a major university. Their algorithm calculated the breadth of the ideas expressed in each essay. It also divided the essays into chunks of about 25 words — without breaking sentences up — and measured how fast the ideas changed from one chunk to the next.
Both broader semantic content and more-gradual transitions between consecutive chunks correlated with higher grade-point averages, the authors found. This was true regardless of factors such as parents’ education level and grades earned during earlier schooling. The correlation was also significant for students whose majors did not involve extensive essay writing, such as engineering, suggesting that certain features of cognitive style revealed through writing can offer advantages in other activities.
