Engagement Doesn't Cause Achievement. It Still Might Be the Most Important Thing in School.

There's an uncomfortable finding that keeps surfacing in education research, and it's one we need to talk about honestly. Student engagement is not always a strong direct predictor of academic achievement.

Rosário et al. (2017) found that while engagement and achievement were correlated, the relationship was weaker than most educators assumed. Carini, Kuh, and Klein (2006) found the correlation between engagement and grades to be weak and not statistically significant. Appleton et al. (2006) found the same for cognitive engagement and academic achievement specifically. And a meta-analysis in Educational Psychology Review (Lei et al., 2025) confirmed that different types of engagement serve different purposes. It found that behavioral engagement was the strongest predictor of achievement, while emotional and agentic engagement served other functions entirely.

So what does all this mean?

Well, a kid can be deeply engaged in your class and not score higher on a standardized test. And a kid who is bored out of their mind might still pull a 95.

If you've been teaching for more than a few years, I’m sure this doesn't surprise you. But it does create a real tension. Because if engagement doesn't reliably move the achievement needle, why should we care about it?

I believe we should care about it more, not less. Here's why.

We've Been Measuring the Wrong Outcomes

The entire framing of "does engagement cause achievement?" assumes that achievement (defined almost always as test scores) is the primary thing school is supposed to produce and care about it. But what if that assumption is the problem?

Gallup's research on K-12 students has consistently found that engagement, hope, and wellbeing account for roughly a third of the variance in what they call "student success", and this is a much broader construct than test scores.

Their data shows that hope is a stronger predictor of college success than GPA, SAT scores, or ACT scores. When more than 900,000 students were polled, only 50% were engaged, 29% were not engaged, and 21% were actively disengaged. And those numbers get worse with every grade level.

That's the "school cliff" many folks talk about, and the finding that engagement drops every year from 5th through 12th grade. If engagement doesn't matter, that drop should be irrelevant…

But every educator reading this knows intuitively that it isn't. It does matter.

What engagement predicts is whether a student cares.

Engagement Keeps Kids Caring About Learning

A longitudinal study of nearly 12,000 French-Canadian high school students (Archambault et al., 2009) found that global engagement reliably predicted school dropout. Among the three dimensions (behavioral, affective, and cognitive) behavioral engagement was the strongest single predictor, which is what we saw above as well.

This aligns with a broader body of research showing that engagement, particularly in its behavioral and relational dimensions, is one of the most consistent predictors we have for whether a student will persist in school, enroll in postsecondary education, and stay through that first critical year. The Check & Connect research program, one of the most rigorously studied dropout prevention interventions, is built entirely on the premise that monitoring and restoring student engagement is the mechanism through which dropout is prevented.

So engagement may not cause a kid to learn more on Tuesday's quiz. But it may be the reason they show up on Wednesday at all. It keeps them caring, and that is huge in my book (should be in yours too).

In a system where roughly 1.2 million students still drop out of high school every year in the United States, that matters enormously.

Why "Engaging" EdTech Often Fails

Now, the skeptics aren't entirely wrong to push back. There's a real problem in education where "engagement" gets conflated with "entertainment," and the research on this is genuinely damning. I’ve written about this “edutainment” a lot, but it’s worth diving into the research one more time.

The seductive details effect is well-documented across multimedia learning research. It shows that adding interesting but irrelevant information to instructional materials (like fun facts, comics, flashy animations) actually hurts learning. Learners process the entertaining details more deeply than the content that matters. Their attention gets diverted. So even if they feel more engaged, they actually learn less.

This is the dark side of the engagement conversation, and it's where a lot of EdTech lives. A platform can be beautifully gamified, full of badges and leaderboards and dopamine loops, and produce students who are clicking constantly but retaining nothing.

The research from Wesenberg et al. (2024) adds an important nuance. These seductive details are most harmful when students are already extrinsically motivated to engage with the material. When motivation is low, those same details can actually help by getting learners in the door.

The lesson here is that performative engagement, you know the kind that looks good on a dashboard but doesn't connect to meaningful cognitive work, is actually a trap.

What Engagement Actually Does (When It's Real)

If engagement doesn't directly cause achievement but does predict persistence, belonging, and long-term outcomes, what's actually happening?

I think engagement functions more as infrastructure. It's not the learning itself, but instead it's the conditions that make learning possible over time.

Consider what we know from adjacent research, that I’ve written about before. Psychological safety (Edmondson) creates the conditions for risk-taking in learning. Autonomy, competence, and relatedness (Ryan & Deci's Self-Determination Theory) fuel intrinsic motivation. Appropriate challenge within a student's zone of proximal development (Vygotsky) sustains productive struggle. None of these directly "cause" a test score to go up on any given day. But without them, learning collapses over the long run (read my post on these conditions here).

Engagement is the felt experience of these conditions being present. When a student is genuinely engaged it usually means that the learning environment is doing several things right simultaneously. The challenge is appropriate, the student feels safe enough to try, the work feels connected to something that matters, and there's a relationship holding it all together.

That matters, and we should keep engagement at the forefront of our work.

Why This Matters More Now Than Ever

Here's where it gets urgent.

We are entering an era where AI can deliver content, answer questions, generate practice problems, provide feedback, and tutor students one-on-one at scale. If the purpose of school is knowledge transfer and test performance, AI is going to be very, very good at that.

So what does school (and human teachers in general) offer that AI doesn't?

It offers engagement in the deeper sense, and maybe the better sense.

The kind that builds identity, agency, belonging, and purpose. Charles Fadelargues that in the age of AI, we need to prioritize the "drivers" of learning: agency, identity, purpose, and motivation. The OECD defines student agency as the capacity and willingness to set goals, reflect on actions, and act responsibly to effect change. These all emerge from environments where students are engaged in the full, multidimensional sense of the word.

When a student is engaged, they are practicing what it feels like to care about something, to struggle with something voluntarily, to be part of a community working toward shared understanding. These are the dispositions that AI cannot replicate and that the future will demand.

If we reduce engagement to a proxy for achievement and then dismiss it when the correlation turns out to be weak, we lose the entire point.

The correlation is weak because engagement isn't doing the same thing as instruction. It's doing something different, and something arguably more important. It's building the kind of human who keeps learning after the test is over.

The Real Question

So no, engagement doesn't directly cause achievement. The research is clear enough on that, and we should stop pretending otherwise.

But the research is equally clear that disengagement predicts dropout, that hope and engagement predict long-term success better than standardized test scores, and that the conditions associated with genuine engagement (like safety, autonomy, belonging, appropriate challenge) are the same conditions that every major learning theory identifies as essential.

"What kind of learning environment do we want to build?"

If the answer is a place where young people develop agency, identity, and the disposition to keep learning in a world where AI handles the routine cognitive work…then engagement isn't a nice-to-have.

It's the whole point.