Campus-Wide Data Reporting and Analysis – Paul Smith’s College
“I don’t have days to spend converting data types or modeling my data.”
Paul Smith’s College had a couple of goals in mind when it began using the Rapid Insight’s Software Suite. They wanted a tool that could combine scattered data from multiple offices and different data systems that don’t always “talk to each other”. The Director of Institutional Research, Dr. Loralyn Taylor, was also the Registrar and held positions on various internal committees, limiting the amount of time she had available to manually format, organize, and prepare data for presentation. She needed to become more efficient and effective with the time and data she had available.
Besides data synthesis and reporting, Paul Smith’s wanted to focus on improving student success and retention rates by using predictive modeling to identify high-risk students at admission, thus more effectively targeting their student support and advising resources. In addition, more accurate admissions forecasting and enrollment management dashboards were also a priority.
I can now transform data into actionable information
To achieve their data reporting goals, Paul Smith’s needed to extract information from various departments including Admissions, the Financial Aid Office, the Academic Success Center and the Registrar’s Office; combining this data meant working with several sets of data definitions and software packages. The individual flaws in each dataset also meant that data integrity was compromised: sometimes a given question would be answered differently by each department asked – something that Paul Smith’s hoped to change.
Rapid Insight’s data preparation tool, Construct, was used to merge these files easily and help fill in any gaps within departments. The outputs of this merge included census files, GPA files, and financial aid files. Each file could be used individually to efficiently create reports, graphs, tables, and dashboards. One feature that Paul Smith’s took advantage of is the ability to auto-email these reports. Instead of compiling data and creating a weekly report to email to board members, Dr. Taylor was able to put the reports on autopilot and spend more time concentrating on solutions. Another benefit was the ability to combine these outputs to create one large dataset that could be used by Rapid Insight’s predictive modeling tool, Predict, to create a predictive model for identifying at-risk students.
Construct, an all-in-one tool for data extraction and reporting, is easily deployed and highly adaptable. Given the wide array of input/output formats and easy to access processing power, Construct blends seamlessly between the nodes of large, complex, and often deeply entrenched institutional data systems.
Predict is a tool for predictive modeling and provides a glimpse into future scenarios based on data found in existing datasets. The statistical software is highly automated, exceedingly flexible, easy to learn, and integrates well with a variety of database solutions and software packages.
We are now able to get answers to our questions quickly and easily
Predict was used to explore statistical relationships between the student’s entry characteristics and their end-of-first-semester GPA to increase retention. It found that several variables, including high school GPA, SAT scores, and chosen major, had a statistically significant effect on the end-of-first-semester GPA.
Through this analysis, Paul Smith’s realized that they were lacking information at the admission stage which could help their modeling. They added questions to the application, as well as considering other information (about roommate, disciplinary history, etc.) to improve successive analyses. They now score each freshman class at admission and evaluate the model’s accuracy at the end of their first semester. The results are used to regularly update both the admission process and the predictive model.
As a direct result of their findings, Paul Smith’s has reorganized from a more traditional Office of Retention to a more holistic Academic Success Center which uses the actionable information generated by the Institutional Research office to be a more effective support center for all students. Predictive modeling and other early alert information is used to accurately identify and prioritize individual students needing additional services. Their findings have sparked campus-wide discussions about student success and how to help those who are not progressing toward graduation including exploring alternative mitigation options, such as requiring summer classes or online tutorials for incoming students who are predicted to be at-risk.
We were able to experience immediate, institution-wide benefits
While using Construct and Predict in the Institutional Research Office, Dr. Taylor found other ways to utilize the program in different departments.
• The Institutional Research Office now has a way to automatically create and email retention reports to executives, as well as managing necessary federal and state reporting. They have also been able to create fixed census files and a campus fact book with information about current students.
• The Registrar’s Office has created a process to find the optimal time to reschedule a class that needs to be moved which includes each student’s and instructor’s schedule, as well as their email addresses to inform them of the change quickly and efficiently.
• The Admissions Office can easily predict incoming class sizes based on the characteristics shared with currently enrolled students. In addition, a detailed Admissions report including predictions for enrollment is automatically generated each week for distribution to faculty and other campus constituents. Critical updates are automatically sent to necessary offices, the VP for Enrollment Management and the President each morning.
Dr. Taylor states that, in the past, this information could only be sent as tables that were hard to read and even harder to interpret. She praises the current Rapid Insight outputs as being both “visual” and “easy for people to understand.”
“We’ve seen significant increases in time savings and efficiency”
The improvements that Dr. Taylor has seen using these programs have resulted in real financial benefits. As for the cost of the programs, she asserts: “If I add up all of my technology fees for Construct and Predict, which is not very much, it literally only takes a few days for me to recoup that entire investment.” Paul Smith’s has seen a return of investment of over $2.5 Million when looking at the students we’ve retained from their efforts over the last two years. Each student retained due to predictive modeling adds to the school’s net revenue. As Taylor states, “You are, in fact, increasing the ability of the college to bring in revenue” by allocating resources wisely.
In addition to a smarter allocation of current resources, Taylor believes that the benefits of predictive modeling have brought in more resources by “increasing actionable information” and allowing the college to ask for more state and federal resources to help them continue to improve their programs.
About Paul Smith’s College
Whether your passion is forestry or the culinary arts, hospitality or biology, natural resources or business, for over 60 years Paul Smith’s College has provided the experiential learning critical to our students’ success. And while our range of associate and bachelor degrees is diverse, everything we teach is based on the same principle: that actually doing is a critical part of learning. Located in the heart of the Adirondack Mountains of New York, Paul Smith’s uses its 14,000 acre setting as a living laboratory to promote economic, social and environmental sustainability.