Improving College Student Success: DataKind’s Continued Work in Post-Secondary Education

By John Harnisher, PhD, Advisor, DataKind Education

DataKind has engaged in two fascinating projects which propose solutions linked to student guidance and help to answer the question which many students pose in their learning journey, “What do I do next?” This question is often an indication of deeper needs or the forewarning of later student disengagement. DataKind has wisely chosen to define direct student success for these projects in terms of more students completing degrees and reduction in time to degree completion.

For the first of these projects, DataKind created a tool, based on historical data, to enable Complete College America to create an optimal schedule for community required college courses. The tool assists community colleges with scheduling classes at more convenient times, so that students complete their courses and, ultimately, their programs at a higher rate.

The second tool, in partnership with John Jay College, generates data-driven insights into which students might be at highest risk of not completing their degree, and recommends when an intervention should take place. This intervention, based on student data, can solve this problem without the heavy resource utilization usually required in such instances by identifying students with the most need. 

DataKind will discuss these two projects at Complete College America’s BREAKTHROUGH conference in Atlanta, GA this week. The annual convening explores innovative policies, perspectives, and practices shaping a more equitable and impactful future for higher education. They’ll discuss the implementation of these tools, impact, and potential to be replicated or scaled. 

This work falls under DataKind’s Economic Resilience portfolio of work and is funded by’s $1.1M grant, enabling DataKind to support a cross-cutting cohort of U.S.-based projects under DataKind’s DataCorps® program.

The next logical step in implementing these tools would be to include more direct recommendations for classes taken based on the steps along a required learning journey and the most effective intervention to keep the students on the learning path. I’m looking forward to the expansion of these projects and inclusion of other volunteers in these very robust research directions. 

College completion and student retention at colleges and universities remains a significant problem. Almost 40% of college students in the US drop out each year ( 2022), and only 41% complete their degrees in four years without interruption. This staggering figure represents a huge loss in terms of time and money spent and loss of potential income.

Two major contributing factors are the lack of clear guidance along the learning journey and a failure to intervene at the right time with at-risk students. Data science has the unique opportunity to mitigate these two problems through the application of good data science practice and the implementation of effective models and tools.

In my experience, learning and education practitioners have made great advances in creating adaptive instructional engines that use learner interactions and data to provide a tailored experience that helps support student mastery. However, a multitude of opportunities exist to make a positive impact on issues that are directly impacting and potentially derailing student achievement. 

There is great additional value expressed in these projects since there was close matching of student outcome goals and organizational success metrics. The tight partnership between DataKind and their partner organizations ensured this direct alignment throughout the projects. 

In my experience, there has too often been a disconnect between the organization’s goals and student success. Even learning organizations can maximize metrics like seat count and lose sight of the longer term metrics like degree completion. A well meaning analysis to increase seats filled could have the result of many classes filled in the middle of weekdays but not allow space for students who need evening and weekend classes. 

Both of the DataKind projects have done an excellent job of keeping all efforts propelling both student achievement and organizational success forward. 


John Harnisher, PhD, is a Learning Science Executive at Educational Initiatives Private Limited with expertise in data science, learning analytics, psychometrics, and learning design.

Header image above courtesy of San Antonio College.

Join the movement.

Quick Links

Scroll to Top