top of page
Search

Data-Driven Insights: Uncovering What Student Voice Data Correlates Most with Persistence




Summary

Capturing and harnessing data to support student success is at the core of recent higher education initiatives. Leaders in student success, such as Dr. KayDee Perry at the University of Olivet and Justin Schuch at Western Illinois University (WIU), have embraced sophisticated analytics and AI tools like EdSights to directly gather and interpret student feedback, effectively listening to student's needs on a large scale. Facing a rise in lagging indicators of student success, such as financial holds and academic flags, these institutions have acknowledged the need to proactively engage with students to identify obstacles to their success as they emerge, rather than waiting for these issues to manifest later on. In collaboration with Dr. Will Miller, AVP for Continuous Improvement and Institutional Performance at Embry-Riddle Aeronautical University, we looked closely at the student voice data that Olivet and WIU are tracking in an attempt to pinpoint the factors that are most impactful on student persistence.


Live Data and Large Sample Size

Two key factors should be focused on when collecting data:


  1. Live Data: Data that is collected and analyzed in real time, with minimal delay

  2. Large Sample Size: Data that represents a large enough portion of your student population 


Often, institutions rely on data from small, potentially biased samples or outdated information that no longer reflects current student demographics or addresses their needs. Platforms, such as EdSights, focus on providing real-time data on student interactions at scale (based on a daily analysis with a 62% average response/engagement rate from students). What sets EdSights apart is not merely its ability to collect direct input from students, but its capacity to analyze this data in real time. This allows EdSights to generate detailed insights, such as each student's risk of not persisting, based on their interactions with the AI.


Initial Findings: Highest Correlation with Persistence Data


  1. Student Change in Risk: The persistence of students who saw a significant risk increase throughout the semester was lower than that of students who were classified as “high risk” entering the term.

  2. Engagement: High risk associated with “Student Engagement” with their institution had the strongest correlation with lack of persistence compared to other categories (Financial, Academic, Wellness). 





Student Change in Risk

Analysis shows that students who move from a low to a high-risk category during a semester are at a substantially increased risk of dropping out—about a 10% higher drop-out rate—compared to those who were classified as "high risk" at the beginning of the term. This increase may be due to the novel challenges and adjustments these students face in their new circumstances, making their situation particularly precarious and urgent. Conversely, students moving from high to low risk show the highest levels of persistence, underscoring the impact of effective interventions at the right time.This data suggests that changes in a student's circumstances throughout the term have a greater impact on their persistence than their initial conditions, even if those initial conditions posed equally significant barriers. Therefore, the timing of interventions is critical, not just their implementation, in supporting students at risk. Moreover, this insight underscores the importance of obtaining real-time feedback directly from students, as the timing of outreach is a key factor in its effectiveness.


Engagement

The data shows that students who self identify as high-risk in terms of engagement encounter the most significant persistence challenges. Feedback from various industry leaders suggests a likely explanation: pinpointing issues related to engagement is significantly harder than identifying academic or financial problems. As a result, engagement issues frequently go unaddressed by conventional interventions. Effectively addressing engagement issues usually demands a proactive outreach approach. Without such a strategy, the lack of student engagement may remain undetected until the student stops out.





Conclusion

The use of AI and data analytics tools like EdSights is transforming higher education by providing real-time, scalable, and systemic insights into student behaviors and needs.  


These insights enable institutions to implement timely interventions and prioritize the students who need their support resources most. This data-driven strategy supports intervention and reshapes the educational landscape through informed, timely actions.


To learn more about EdSights and how we engage students, please email nick@edsights.io.

75 views

Comentarios


bottom of page