Skip To Content

Why Law Schools Should Consider Predictive Modeling

Reading time: 3 minutes
Law School Data

By Caitlin Garrett, Statistical Analyst

It’s no secret that the law school environment has gone through some dramatic changes in the past few years. For me, this realization came while helping one of our law school customers with a predictive model they were building. No matter which way we sliced their data, it seemed there were some external factors that had changed their student population in a matter of a few years. After doing a little bit of research, we found the below graph, and noticed the steep drop in the number of administered LSATS nationwide.


You might imagine that fewer LSATs administered relates very directly to the number of students who apply to and enroll in law schools.  And you’d be right. Currently, first year enrollment is at levels unseen since 1977, making this incoming class the smallest in 38 years. Many colleges are trying to attract applicants by waiving application fees, so students are applying to more law schools. Factor in the fact that 85% of law school grads have loan debt nearing 100k and that the job outlook hasn’t bounced back from pre-recession levels (and isn’t expected to), and we have a kind of perfect storm of reduced candidates for law school entry. Within this environment, it’s more important than ever to keep on top of enrollment numbers.

All of this has resonated with our customers’ experience. Law schools, as a whole, have been quick to adapt and innovate whenever possible. With fewer tuition dollars, many have cut staff and trimmed faculty in order to balance budgets. For state schools, reducing the amount of tuition paid by out-of-state students has helped to attract students who might not have previously considered them. In order to keep up class size, some have relaxed admissions standards, while others have intentionally shrunk the number of admitted students to preserve their standing. While there are many ways to tackle enrollment issues, I’d advocate for using data to drive decisions. What we’ve seen through our customers is a delicate balancing act between focusing on enrolling quality students and managing expenses. Rather than relying on their old methods, they’ve employed predictive modeling to inform their enrollment decisions.

So how are law schools using predictive modeling? By using historical data about who enrolled in the past (and who didn’t), schools can predict student by student enrollment likelihood. By combining the resulting individual enrollment probabilities, schools can get a good idea of what their next class will look like – including yield, seats filled, and academic indicators like LSAT median and quartile. From there, schools can create an estimated financial aid outlay by using financial aid information along with predicted enrollment scores and use predictive modeling to help allocate financial aid packages. Besides enrollment, they can predict who will pass the bar on the first try, who will be retained, and even after graduation, who is likely to financially give back to the program.

While predictive modeling might be new to law schools, it’s been around for a long time. Companies like Amazon, Netflix, McDonald’s, and Target all use predictive modeling every day to optimize their resources. It took a little longer to adopt predictive modeling in the higher education world but it’s had a powerful impact for many undergraduate programs. One thing we’ve learned through our customers is that many law schools don’t have a statistician on hand – and that’s okay. If you’re considering modeling as an option, pick a tool and a knowledge partner who does have that statistical expertise and who you can learn from.

Notify of
Inline Feedbacks
View all comments