<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=341153139571737&amp;ev=PageView&amp;noscript=1">
Back to all

Data Analysis Holds Key to Predictive Modeling

For years, predictive modeling has been used by a wide range of companies and organizations to determine the statistical probability that an event will occur. Historically, the process has been embraced primarily by businesses, insurance agencies, and health professionals.

Now predictive modeling has become a useful tool in law enforcement, archaeology - and educational institutions.

Higher Ed is increasingly embracing the concept. Most colleges and universities are being asked to do more with fewer resources.

Improvements in student retention, degree completion rates, and support services are among the priorities that institutions are up against. Predictive modeling can help institutions allocate resources to provide the most assistance for attaining their goals whether these goals involve creating a better experience for underserved populations, increasing donations from alumni, or improving completion rates.

Central to predictive modeling for Higher Ed is the principle that both successes and failures leave clues to help identify the causes.

Much like a forensic scientist, institutions can analyze these clues to discover commonalities. For example, what traits are shared by students who fail to return after a semester ends? What traits are shared by those who drop courses in the middle of the semester? What are the common attributes among students who persist until graduation?

The variables that need to be included depend on the outcome that you need to predict. Here are just a few of the variables that can be valuable for predictive modeling:

  • Student age at enrollment
  • Ethnicity of student
  • Distance between student's residence and campus
  • PLA credits earned
  • Number of classes taken in the semester
  • Difficulty of classes taken
  • Student's employment status
  • Source of student's funding for tuition
  • Declared or undeclared major
  • Number of classes dropped
  • Student's education level at enrollment
  • Marital status
  • Grade point average
  • Instances of academic probation
  • Campus-related extracurricular activities
  • Income of student and/or student's parents
  • Remedial courses needed or taken
  • SAT or ACT scores, if available

Based on the data collected, you can analyze information to answer many questions. Students in which majors are most likely to complete their degrees? How does financial aid affect retention? What impact does PLA have on completion rates? How does distance affect retention and/or completion? Are donations from alumni affected by major? Is age, ethnicity, or income the better indicator of student success? How does spending on support services affect retention or completion rates? Is there a correlation between the number of credits transferred from another institution and the student's persistence? What are the average grades for students who earned a GED rather than a high school diploma?

Having identified trends, you can ensure that you allocate the appropriate resources to support students who are at the highest risk of attrition.

Perhaps you discover that first-generation students possessing a GED have a higher attrition rate if they attend full time during their first semester. Incoming students with these traits could be advised to take lighter loads for the first semester or two.

Your data might show that students living more than 30 miles from campus tend to drop classes at a higher rate. If you have a number of students with long commutes, you could consider expanding your online offerings.

Historically, institutions have tried to use SAT/ACT scores, high school grades, and similar metrics to predict a student's success. However, there are many college drop-outs who had respectable test scores and grades, and there are many students who barely met the minimum qualifications who earned their degrees with impressive GPAs.

Obviously, other factors contribute to student success or failure, and those factors can be found in amassed data. Predictive modeling can help those who offer learning opportunities to direct efforts more precisely to serve students while simultaneously achieving goals.



Want to learn more? Let's talk!

If your institution or organization is seeking guidance, solutions, or support from CAEL, or if you have an idea for a future collaboration initiative with us, please reach out. We'd love to connect.

Contact Us