Using Analytics to Improve Student Success

Mathematics is a powerful tool.  We use applications from math all day, every day, like when we drive over a bridge or look at our email.  

Math is very difficult for many people, and they struggle with it.  Many students at our university (and almost all others) need to repeat a math class to progress in their program and complete their degrees.

Our research has uncovered some important facts:

• Many students repeating math classes often must do so more than one time, which extends their time to degree and could prevent them from graduating.  

• Students who use the Math and Science Tutoring Center improve their chances of passing their math courses.

• The positive effect from math tutoring is especially large for students who are repeating a course.  “Repeaters” increase their chances of success by roughly 15 to 20 percentage points!

• But, we found, only about 10 percent of students repeating math were going to tutoring

The Data Science Team, supported by DoIT and the Provost's Office, designed a pilot program in conjunction with the Math Department to “nudge” students repeating 100-level math courses to use tutoring.  

A nudge is a message designed to provide information to many people at very low cost that changes behavior over time.  An example of a nudge you see in everyday life is the roadside sign that says “Click it or ticket.”  Our nudges let students know not just that tutoring is effective, but also how effective it is.

This semester (Spring 2021) we sent nudges to 114 students using a myUMBC personal post.  About 70 percent of students receiving the nudge read them, and we have evidence that suggests that the nudge leads to an increased use of math tutoring by roughly 30 percent.  

The pilot is now in its 8th semester.  During that time we’ve nudged over 1800 students repeating their courses.  

In addition to this program, we’re using what we’ve learned about nudges to help students in other ways that support their success, including combining them with predictive analytics for enhanced early alerts.

By Bob Carpenter