4 Insights from Engaging Student Mindsets at Michigan Virtual

In this guest blog, Dr. Nikolas McGehee, Senior Data Scientist at Michigan Virtual shares his team’s insights from partnering with Discourse Analytics. Any institution looking to implement technology and AI tools like DA’s Digital Counselor can benefit from what Michigan Virtual learned. 

At Michigan Virtual, we aim to Drive Educational Change and Broaden Student Pathways. To help us reach that goal, we partnered with Discourse Analytics to gain insights into our students’ mindsets and behaviors utilizing the power of Artificial Intelligence and Machine Learning. Discourse Analytics has played an important role in enhancing student outcomes and providing personalized support at Michigan Virtual.

How Michigan Virtual Engaged Discourse Analytics

Discourse Analytics’ platform analyzes student behavioral data to create mindset profiles of students. Mindset profiles represent how a student views the world around them during a period of time. Importantly, Discourse Analytics doesn’t use demographic data from our institutional dataset. We chose to partner with Discourse Analytics to help us understand our students more deeply and provide proactive, personalized support.

1. The Value of Attitudinal Dimensions

One of the most significant insights we gained from working with Discourse Analytics was the importance of attitudinal dimensions in predicting student achievement. Attitudinal dimensions are characteristics that make up a student’s mindset profile. In essence, those mindset profiles show how students may perceive the world around them.

We found that certain attitudinal dimensions were particularly predictive of students’ success. Resiliency, focus capacity, technology, proactivity, and motivation were especially powerful indicators of student achievement. By identifying students with low scores in these attitudinal dimensions, we were able to implement targeted interventions to support their academic growth.

2. The Nuance of Attrition Risk

Discourse Analytics also provides insight on student attrition risk. At Michigan Virtual, we chose to view this metric as a scale, rather than as a binary indicator as Discourse Analytics does by default. This offered a more nuanced perspective of student attrition risk.

By examining a student’s attrition score and how it changed over time, we identified two types of students who might otherwise gone unnoticed: those who were struggling academically but were still making progress and those who were meeting quality indicators but not submitting enough assignments.

Through this approach we identified students who were at risk of attrition but who might not have been flagged by a simple binary classification.

3. What We Learned About Nudges

We leveraged DA’s mindset profiles and risk flags to nudge students to success. Nudges are timely reminders and encouragement meant to help students progress in their classes. In our pilot with DA, we explored the effectiveness of teacher-initiated nudges based on student profiles. We gained a couple of great insights about nudging through our work with Discourse Analytics.

The first valuable lesson we learned from our analysis was the effectiveness of early intervention. Students who were struggling to keep pace were more likely to pass if they engaged with nudges and their resources in the first 10 weeks of the semester. Early intervention was more impactful.

Our analysis also revealed that different types of nudges had varying levels of effectiveness based on students’ recommended engagement strategy. An engagement strategy is a series of messages proactively addressing an area a student could struggle in. Personal wellbeing and academic nudges were particularly impactful, regardless of the student’s mindset profile. We encountered some challenges in implementing a system that would deliver targeted nudges at scale during this pilot and are exploring how we can grow this strategy.

Finally, we learned that writing the nudges ourselves was most effective. We wrote the nudges because we know our students best. That way we were able to keep our voice while addressing student challenges as identified by DA’s platform.

4. The Importance of Data Expertise

Discourse Analytics provided valuable insights, but we quickly realized the importance of having a data scientist on our team. To fully leverage the tool’s potential, we needed someone who could integrate the data with our existing systems and extract meaningful information.

This was key for us and is something to consider for any educational institution using data and analytics to make their investment worth it. At Michigan Virtual, we are very fortunate to have data scientists on our team and recognize that such expertise is rare in K-12 institutions.

Lessons Learned and Future Directions

Our experience with Discourse Analytics has been valuable in gaining a deeper understanding of our students and improving our educational practices. We’ve learned the value of attitudinal dimensions, the effectiveness of early intervention, and the importance of data expertise.

With the dashboard, we were empowered to look at what our students were struggling with and how we could help them. As we move forward, we are committed to further refining our use of analytics to enhance student success.

Connect with us to learn more about how you can engage with students’ mindsets.