Learning Moments: Illume® Insights

Civitas Learning Civitas Learning Space Data Science Results ROI

Share this Post

As part of our promise to build a community that is Learning Together®, Civitas Learning and our partner institutions have shared insights via the Austin Community College Learning BriefUniversity of Arizona Learning Brief, Three Case Studies to Consider in the RPA Journal, and the UMUC Case Study. In addition to that, we intend to share periodic insights important to the collective community. Recent work with our partner institutions using the powerful Illume application have surfaced important insights and knowledge that can benefit our community of practice.


The core purpose of the Illume app is to turn the lights on for administrative staff and decision-makers at a particular institution. Users are able to analyze key institutional metrics, starting with persistence (a student’s likelihood to remain enrolled at the school) and identifying key institutional insights that could be used for policy and decision-making. A few examples of what that means in terms of actual learnings follow below.
Learning Moment #1: Persistence and GPA Tipping Points
Background One of our Pioneer Partner institutions was working with external consultants and national research metrics to determine which group of their students were at risk of not persisting and were in need of advising interventions. Based on that general research, the college was advised that students with a GPA of 2.2 or below were typically those that need help. They then targeted this particular group of students with a robust intervention program which had an overall increase in persistence that was less than expected. Insight Once this college utilized our platform and deployed the Illume application, they discovered that for their college, the GPA tipping point (where the likelihood of persistence starts to switch to being more likely to persist from being less likely) was not at 2.2 as the general research and external consultants indicated, but was in fact at 2.6. Their program was essentially missing a critical group of students (comprising 25 percent of the students at-risk) who were more likely to succeed with the right nudge. Action This insight led to the college expanding and testing interventions to include outreach to students within the specific at-risk GPA range for their college. gpa tipping point graph
Learning Moment #2: Enrollment Persistence and Scholarships
Background This Pioneer Partner institution had an enrollment target of 10,000 students following a four-year decline in student enrollment. In order to draw more students, they had been awarding attractive scholarships for first-year students, based on each student’s ACT scores – a fairly common practice among institutions. Insight Once our partner implemented our platform and deployed the Illume application, they discovered that a number of students given priority for admissions and scholarships based upon their ACT scores were transferring out to other institutions prior to earning credentials — using this institution as a stepping stone. More importantly, Illume insights uncovered that at this institution, high school class percentile was a better indicator of persistence for students than ACT scores. Action Today, the institution is re-evaluating their admissions and financial aid policy to make changes in the way they evaluate student applications and allocate valuable scholarship dollars in order to hit their target enrollment rates. persistence predictors
paired predictor plots

The above two representations of the same data show that High School Percentile is a better indicator of student persistence when compared to ACT Score at this institution. Moreover, data show that leveraging a combination of both predictive features is an even better indicator of student success. The heat map on the right further shows the highest concentration of persisting students in this quadrant.

Learning Moment #3: Identifying Pockets for Improved Retention
Background This university has done extensive work in analytics to identify factors that help predict student success including, but not limited to, academic readiness, residency, and gender. An enrollment projection model developed by their Institutional Research team, is a model that is used and maintained by the university to predict first-year GPA for incoming students. However, while this model provides valuable information for planning decisions, it does not reflect the rich data that become available after a student arrives. Insight Using the Illume application, the university uncovered 660 non-resident female students with GPAs between 2.4 and 3.0 who were predicted to be at risk of not persisting. This insight identified an important segment that may have been overlooked due to the lower risk associated with this segment in best practice research based on gender and GPA. Action The university is now working on interventions to the group of 660 students identified through Illume. Their integrated approach includes timely and appropriate emails, nudges, phone calls and additional information to help personalize the right outreach to get, and keep, these students on track to persist. Outreach is being tested to determine the efficacy of each intervention.