Finding the Belonging in the Data: How Sentiment Analysis and Student Stories Are Transforming Illinois State University
Jayla Langford, NSSE Institute Graduate Assistant
At Illinois State University (ISU) located in Normal, Illinois, NSSE data isn’t just about numbers. It’s about stories—especially those that illuminate how students feel about where they are and who they’re becoming.
“We’re discovering that students are looking for relationships,” said Dr. Ryan Smith, The Director of University Assessment Services and faculty member at ISU. “The students are inundated with content. But it’s relationships and interactions with faculty that blow everything else out of the water when it comes to retention.”
ISU administers NSSE every three years to unpack what makes students stay and thrive. In a recent wave of innovation, assessment offices are merging quantitative and qualitative data to dig deeper into student experiences.
In 2023, about 50% of ISU’s NSSE respondents answered the open-ended, qualitative questions. With more student voices in play, Smith’s office used sentiment analysis tools in Power BI to make sense of the emotional tones buried in hundreds of written responses (Smith, 2024a). His qualitative curiosity sparked a shift in how assessment units at ISU understand the student experience—from raw numbers to raw emotion.
Institutions like ISU are turning to the open-ended, narrative responses from students to shape more human-centered strategies. For assessment professionals at ISU, sentiment analysis—a method of using AI tools to detect emotional tone in text—is proving to be a transformative bridge between data and narratives. Sentiment analysis, paired with predictive modeling, is helping see the full picture—not just who is struggling, but why.
The initiative is grounded in a broader effort in assessment at ISU to explore student retention (Smith, 2024b) through the lens of belonging. “Belonging was everything in our retention model—far and away,” Smith mentioned, referencing the assessment unit’s ongoing analysis using NSSE data. “GPA is always the highest, but I think GPA is a signal of belonging. It’s not just a predictor, it’s an outcome.”
Another trend Smith’s office found from the NSSE data suggest that while some groups—like LGBTQ+ students—feel more comfortable being themselves over time, they also report feeling less valued by the institution by their senior year. “We found that LGBTQ+ students reported feeling less valued by the institution over time,” Smith said. “But their sense of being able to be themselves actually increased.” It’s a paradox that only emerges when data is disaggregated—and when qualitative responses provide the context.
For Smith, these insights underscore the need for systemic change in changing assumptions about how – and where – students learn. “Learning happens everywhere,” he said. “It doesn’t just happen in the classroom. It happens in student organizations, in undergraduate research, and in conversations with faculty. The most successful institutions in the future will be the ones positioned to embrace that.”
Smith’s growing embrace of qualitative methods has been transformative for his work in student success and engagement. Though trained in quantitative analysis, he has come to believe in the power of narrative. “I’m quantitative by training, qualitative by nature,” he said with a laugh. “And what I’ve learned is that our leaders say they want metrics, but what they really need are stories and decisions. Belonging, once a vague concept, is now a measurable force at ISU. And it doesn’t look the same for everyone stories.” That approach has led him down a path of deep inquiry into student voices—particularly how students express belonging.
Such storytelling adds validity to institutional data. “When I do quantitative analysis, and then I see it verified in a student’s own words—it’s powerful. Quantitative analyses aren’t always telling the story. When the audience is seeing it in the students’ own words, it’s powerful. But I really think if one wants to tell our story, we really need to focus on the qualitative aspect of it.”
For years, Smith and his team have built extensive dashboards to visualize NSSE data. But he noticed something: while charts could reveal patterns, quotes made people care. ISU prides itself on being a relationship-rich institution. According to Smith, that culture of connection plays a key role in assessment as well. He works closely with Dr. Erin Thomas in Student Affairs to ensure that academic and co-curricular data aren’t siloed. Their collaboration has led to presentations across campus, student affairs staff understand and act on NSSE findings.
This has led to what Smith calls “data storytelling,” a practice that helps bridge the gap between abstract numbers and lived student experiences. He recalled presenting a chart on Hispanic student belonging—powerful data in its own right—but what truly resonated was a quote from a Hispanic student describing their lived experience at a predominantly white institution. “I showed a chart with Hispanic students, and it was compelling,” he said. “But then I shared a quote from a student about her experience at a predominantly white institution. That’s what the audience latched onto.”
That realization is reshaping how assessment units in academic and student affairs approach retention.
This narrative approach is also shaping internal conversations at ISU. While presenting NSSE data to faculty and staff, Smith has noticed a divide: faculty often focus on methodology, while student affairs professionals dive straight into the stories. Both views are necessary, he said—but leadership must ultimately focus on impact.
Still, changing institutional culture isn’t easy. Smith describes ISU as a relationship-rich campus, yet he admits the challenge lies in leveraging that strength with data. “It’s difficult to prioritize using evidence for student success,” he mentioned. “It takes a cultural change—and we’re not there yet.”
Smith is hopeful for their 2026 NSSE administration. He envisions a more streamlined, targeted approach to the survey itself and a deeper institutional commitment to using its findings. “If I could do one cycle on belonging and interactions, another on time use, another on faculty engagement—that would be ideal,” he said. “Because belonging is still the number one factor.”
References
Smith, R. (2024a). AI sentiment analysis of survey text responses: Not too shabby. Illinois State University – University Assessment Services Projects. Retrieved from https://about.illinoisstate.edu/uasprojects/2024/09/10/ai-sentiment-analysis-of-survey-text-responses-not-too-shabby/
Smith, R. (2024b). Can I do it? Do I belong? Should I do it?* Analyzing retention at ISU using the National Survey of Student Engagement. Illinois State University – University Assessment Services Projects. Retrieved from https://about.illinoisstate.edu/uasprojects/2024/03/12/can-i-do-it-do-i-belong-should-i-do-it-analyzing-retention-at-isu-using-the-nsse-national-survey-of-student-engagement/
