Student voice improves outcomes when students can shape teaching
Updated Jul 23, 2026
A university can report solid engagement and still miss a more important question: do students think they can influence what happens around them? At Student Voice AI, we see that gap in open comments all the time. Students often describe whether staff listened, whether they could raise concerns, and whether speaking up felt worthwhile. That is why Uriel Eduardo Torres Castro and Clelia Pineda-Baez's Higher Education paper, "Student agency in Colombian higher education: a dual-pathway model of mediation and moderation between student engagement and academic achievement", matters for universities using student voice to improve teaching and student experience. The paper argues that engagement becomes more educationally useful when students can shape the learning conditions around them, not just participate within them.
Context and research question
The paper begins with a problem many UK universities will recognise. Quality assurance systems are good at counting outcomes, participation, and performance, but much weaker at showing how learning is actually experienced. The authors argue that educational quality is often flattened into metrics when it should also be understood through the relational processes that help students ask questions, influence teaching, and turn participation into progress.
To test that claim, the study analysed data from 1,713 final-year students across six accredited private universities in Colombia. Using National Survey of Student Engagement indicators, the Agentic Engagement Scale, hierarchical regression, and structural equation modelling, the authors asked whether student agency, expressed through agentic engagement, both mediates and moderates the relationship between engagement and academic achievement. For UK Student Experience, Quality, and Market Insights teams, that makes the paper useful because it turns a familiar idea, that student voice matters, into a more precise question about when it actually changes outcomes.
Key findings
Collaborative learning and student-faculty interaction remained the two engagement indicators most clearly associated with academic achievement. When the study modelled the full set of NSSE engagement indicators together, those two relational forms of engagement still showed unique, statistically significant links with GPA. The practical message is straightforward: the kinds of engagement most likely to matter are the ones built through interaction, not only through individual effort or simple participation counts.
The key agentic mechanism was expressive voice. Among the five agentic engagement items, the one that mattered was the simplest: whether students expressed their preferences and opinions in class. That item mediated the relationship between both collaborative learning and student-faculty interaction on one hand, and achievement on the other. The indirect effects were modest, but they were statistically significant, and about 6% of the total effect of collaborative learning on GPA was channelled through expressive voice. For UK teams, the takeaway is that student voice is not a soft extra. It is one of the ways engagement becomes educationally consequential.
The paper's wider framing is also useful:
"foregrounding how learning is enacted rather than merely reported"
Expressive voice also changed how far staff interaction needed to carry the load. The moderation analysis showed that student-faculty interaction had a stronger relationship with GPA among students with lower levels of expressive voice, while the relationship was still positive but less pronounced for students who were already more agentic. In other words, students who can speak up and shape the learning process appear to be less dependent on intensive staff interaction for academic success. That does not make staff contact less important. It means voice partly compensates when other relational opportunities are thinner.
The broader implication is that educational quality should not be reduced to whether students are present, busy, or satisfied. The paper treats student agency as a situated mechanism that helps explain why similar levels of engagement do not always lead to similar outcomes. For universities, that matters because the same attendance pattern or survey score can sit on top of very different learning conditions. If students do not feel able to question, suggest, or influence, engagement can look healthier than it really is.
Practical implications
For UK universities, the first implication is to treat the ability to speak up as part of educational quality, not just as a classroom personality trait. Add prompts to local surveys, module evaluations, or pulse checks that ask whether students felt able to express preferences, ask questions, or suggest changes during the course. The benefit is earlier evidence about whether teaching environments support active participation or quietly suppress it.
Second, institutions should separate participation from influence when analysing student experience evidence. A dashboard can show attendance, platform activity, or response rates, but it cannot show whether students felt heard. Pair headline measures with open-text prompts and a reproducible method such as the NSS open-text analysis methodology so recurring themes about staff responsiveness, collaborative learning, and confidence to speak can be compared properly. That gives teams a cleaner diagnosis of where student voice is supporting learning and where it is only being counted.
Third, universities should build more than one route for agentic participation. Some students will speak in class. Others will only raise a preference, confusion, or barrier through written feedback, low-stakes check-ins, or asynchronous discussion. If institutions want student voice to reach beyond the most confident students, they need routes that do not depend on confidence alone, and a clear student comment analysis governance checklist if those routes generate sensitive qualitative evidence. The benefit is a more inclusive evidence base and more targeted follow-through.
Student Voice Analytics fits naturally here because it helps universities group open comments about teaching interaction, support, collaboration, and whether students felt able to influence the learning environment. That makes it easier to see patterns across modules and cohorts, rather than relying on a few vivid remarks or a single aggregate score.
FAQ
Q: How should a UK university apply this paper when reviewing student engagement or student voice?
A: Start by checking whether your evidence captures influence as well as participation. Ask students whether they could question, suggest, or clarify during teaching, not only whether they attended or were satisfied. Then review those answers alongside open comments to see whether voice is strongest in some programmes, disciplines, or delivery modes and weaker in others.
Q: What are the methodological limits of this study?
A: The study is based on final-year students at six accredited private universities in Colombia, and it uses statistical modelling rather than experimental or longitudinal design. That makes it strong evidence about relationships and mechanisms, but not proof of causality, and the effect sizes for the indirect pathways were modest. UK teams should transfer the underlying logic carefully, then test the same questions in their own settings.
Q: What does this change about student voice work more broadly?
A: It shifts the focus from whether students participated to whether they could shape the conditions of participation. Student voice becomes more useful when institutions ask not only what students did, but whether they felt able to influence teaching, ask for clarification, or make their perspective count. That is the point at which open comments stop being anecdotal and start becoming evidence about educational quality.
References
[Paper Source]: Uriel Eduardo Torres Castro, Clelia Pineda-Baez "Student agency in Colombian higher education: a dual-pathway model of mediation and moderation between student engagement and academic achievement" DOI: 10.1007/s10734-026-01669-3
Request a walkthrough
Book a free Student Voice Analytics demo
See all-comment coverage, sector benchmarks, and reporting designed for OfS quality and NSS requirements.
- All-comment coverage with HE-tuned taxonomy and sentiment.
- Versioned outputs with TEF-ready reporting.
- Benchmarks and BI-ready exports for boards and Senate.
UK-hosted · No public LLM APIs · Same-day turnaround