I'm Not A Statistician, But I Play One On TV: A Retention Analysis
Concurrent Session 5
This discovery session will walk thru a workflow that the IU Online office uses to evaluate retention. We will touch on the different factors we are using to calculate retention, our future methods for predicting retention, and the tool that allows for a non-statistician like me to build predictive models.
Understanding the factors that impact retention for students in online programs is essential. In addition to student demographics, we are currently looking at whether or not targeted messaging via a third party vendor, as well as student support services, have any impact on student retention. Because this data, along with our student information data, comes from different data sources, it can sometimes be difficult to bring all of it together. However, with the right tools, bringing this type of data together doesn’t have to be so pain-staking. Our goal for this discussion is to introduce how we leverage Alteryx to study retention, and look at how this tool can also preform predictive modeling to determine the key factors for retaining a student.
This discovery session will be a live demonstration of our methods within Alteryx. We will provide a high level view of the workflow we use to explore retention rates. This presentation will provide an up close look at the tool, and demonstrate a few of its features that allow you to clean and blend data from several disparate data sources. We will also briefly discuss how we plan on using the tool in the future to help us build a predictive model for retention. By demonstrating the tool via a live demo, we will provide insights into our methods for studying retention, and we will give attendees an opportunity to get up close and personal with a new analytics tool.
The main takeaways from our presentation include:
- Insights into our retention calculation efforts
- How leveraging a new tool can turn a complex problem into a simple solution
- Newer tools are out there to help non-statisticians better understand their data