Q&A: Admissions Modeling

In preparation for his presentation at the Rapid Insight User Conference, I got the chance to sit down with John Keyser to ask him a few questions about admissions modeling.

It seems like law schools have really taken an interest in predictive modeling lately. Is this a new thing for them? Are other law schools doing it?

I think it’s pretty unusual. Most law schools have not had institutional research arm – they have relied on the larger campus. They have not been equipped to use analytics. Generally they hire lawyers as administration, so it is rare that they are using these techniques. I think about five years or so ago, some schools were being pressured to be more analytical as application numbers declined and the financial picture of law school changed. Others might have heard through the grapevine that law schools are doing it and feel like they’re falling behind. My background is as a quantitative social scientist, so I brought knowledge of multivariate analysis to the job. I am serving on a new steering committee of Associate Deans of Finance and Administration and there is a movement that is starting to organize to bring a more analytical approach to law schools.

You mentioned in your webinar blurb that admissions modeling is both an art and a science. What does that intersection look like for you?

I think non-analytic people have a fear that quantoids are going to move in and take over their turf, and nothing could be further from the truth. I think these tools allow us to provide some benchmarking on what’s working and what’s not working. They also tell us which variables are really relevant in the enrollment process. It’s still a lot of gut instinct, so the admissions staff is a crucial part of the team when analyzing data. You have to have a theoretical underpinning to the models and that’s where their knowledge is crucial. Modeling is a way to help quantify the information, and while I do the quantitative part, I value the opinions of other professionals in that area.

What advice would you have for someone getting started with modeling?

Make sure that any data being pulled is actually the data you need and that you have a good working relationship with the office you’re working with. Sometimes you won’t have access to raw data and someone else will have to pull it for you. You have to make sure they give it to you in a consistent way. Work with them to figure out targets. For us, we’re trying to hit median LSAT and GPA scores – make sure everyone understands the targets and the types of categories you’ll put people in. Define the process up front. You want to shoot for a seamless handoff of really good data to analyze.

How similar is modeling for a law school to modeling for other colleges or grad schools?

I think almost everything in my presentation will be applicable for other colleges. I’m using the law school context, but it applies to any process that attempts to model the admissions process. The admissions processes aren’t the same but they are very similar.

Decentralize analytics.
Harness the power of many.

Create and share reports and datasets across the enterprise, and put analytical power in the hands of everyone. Veera creates a truly data-driven culture. Try it for yourself today.



Decentralize analytics. Harness the power of many.