Enrollment Projections: The Pig in the Python

See this session and others at AACRAO's SEM Conference in Los Angeles, CA.  Register today!


Luke  Schultheis

Vice Provost for Strategic Enrollment Management at Virginia Commonwealth University

Alan Sack

Data Analyst at Virginia Commonwealth University

Enrollment Projections: The Pig in the Python

Tuesday, October 28, 2014 11:00 AM - 12:30 PM


Can you elaborate on this ‘Pig and Python’ enrollment projection metaphor?

Alan: To start with, I have been doing institutional research for over 25 years.  While working with the director of IR at the time, we started noting that an increase in new student enrollment has a ripple effect for at least 4 years. People couldn’t see why there was an increase in enrollment after the initial surge. The reason is that cohort stays with you for a number of years. So the easiest way to think of it, visually, is to think of a pig in a python. This is important, because when you show that enrollment increase from year to year after the surge, some people are inclined to think, ‘oh, we must have had a really good retention rate, or a good transfer class,’ when really, the majority of that percentage increase comes from that initial surge. So we make our models based on that cohort.

What does this mean for the rest of the institution?  

Luke: Identifying and communicating this projection is very important for the university. Housing and business services need to forecast or project their potential costs. Without understanding that this cohort will be staying with the institution, forecasts on the number of beds or available dorm rooms might be made based on the number of incoming students. Need to look at cohort size.

Conversely, when you have a dip in incoming new enrollments, it works the other way. Senior administrators often think that by simply increasing the number of freshman after an enrollment dip, you will have solved the problem, but you still have a small cohort ahead of that class which needs to be factored in to departmental forecasts.  

I won’t lie – I am no statistician. Can you elaborate a bit more on the model?

Alan: The model is looking at the impact of a surge of incoming people, and then factoring in their progress through the student life-cycle: freshmen, freshmen to sophomore, sophomore to junior, junior to senior, etc. As I said, I have been using this model for a long time, and it is pretty easy to swallow. If people don’t understand it, I can throw out the pig and the python and it is easy to visualize. There are some cutting-edge techniques used by many IR professionals for predictive modeling, like using Markov Chains and Multiple Regression. They do an excellent job as well, but the data from there may be a bit harder to communicate to the university.

Luke: Communicating these projections is very important. Some institutions with a sophisticated enrollment management model might have an idea, but many operations are disjointed – not lots of communication between departments. At least in my experience, the IR office doesn’t look at nuances of institutional operations so it is important to bring them in so they can produce data that is more usable for the entire university.

It has to be fundamental across the institutions well. If changes are not communicated they will affect the institution in unforeseen ways.

Do you have anything else you would like to add?

Luke:  Our office are adjacent to each other. While I preach communication, Alan and I have disparate tastes in music and typically raise our volume to drown out the other’s inferior choice.

Alan: We crank it up.