Classroom Practice · CLSM Framework
The Most Engaged Student in the Room May Not Be the One Learning the Most
One of the strangest things about teaching is that I can see almost everything a student is doing and still have very little idea what is happening in their mind.
I can see the Chromebook open. I can hear the discussion. I can watch someone highlight a reading, fill in notes, answer questions, move around the room, make a presentation and stay busy for an entire class period. If I walked past that student twenty years ago, I probably would have thought I was looking at engagement. I still might. What I am less certain about now is how much that tells me about learning.
That distinction has been bothering me lately because engagement has become one of those words in education that seems obvious until you try to define exactly what you mean by it.
Researchers don't necessarily treat engagement as one thing. Behavioral engagement can involve participation, attention and persistence. Emotional engagement gets closer to interest, enjoyment or a student's feelings about the learning. Cognitive engagement deals with something harder to observe: the mental effort being invested in understanding, connecting, reasoning, monitoring and making sense of information. Recent research continues to separate these dimensions because they do not always move together.
That makes classroom observation much more complicated than it first appears.
A student can be behaviorally engaged while doing relatively little with the content itself. Another student can look fairly passive while trying to connect a new idea to something learned yesterday. A conversation can involve deep reasoning or almost none. Highlighting can require a student to make decisions about importance, or it can become something the hand does while the mind is somewhere else.
The behavior doesn't tell the whole story.
There is research behind this problem that I wish I had understood earlier in my career. The ICAP framework, developed by Michelene Chi and Ruth Wylie, distinguishes passive, active, constructive and interactive forms of cognitive engagement. The interesting part is not simply that students should “do things.” The type of thinking elicited by the activity matters. Generating an explanation, making an inference or building on another person's reasoning asks something different of the mind than copying or manipulating information already provided.
Newer studies are still finding useful differences along those lines. A 2025 experiment involving instructional video found differences in cognitive load, retention and learning depending on the kind of interaction students experienced, with richer forms of interaction producing broader benefits.
I find that interesting because classrooms are full of activity now.
Students drag things. Click things. Create things. Discuss things. Respond to polls. Make slides. Watch videos with embedded questions. Collaborate on documents. Rotate through stations. Highlight digitally. Move information into boxes.
Some of those things can produce excellent learning.
Some can produce excellent-looking classrooms.
Those are not guaranteed to be the same classroom.
I have become much more interested in the cognitive work hiding underneath an activity. If students are making a presentation about an empire, for example, I want to know where the decisions are. Did I already tell them which information matters and where to put it? Are they mostly transferring information from one place into another? Or do they have to decide which developments mattered, connect political decisions to economic consequences, defend why one factor was more significant than another, or notice a relationship I never explicitly gave them?
The presentations might look almost identical when they're finished.
The thinking required to produce them could be completely different.
That gets very close to something I've been trying to understand through CLSM.
Cognitive load matters because students have limited capacity to work with information at one time. Language matters because students cannot reason deeply about information they cannot adequately access. Structure matters because organization can free attention for the content instead of making students spend it figuring out what to do. Memory matters because whatever happens during the lesson eventually has to survive beyond the lesson.
But none of those pieces tells me much if I never ask what the student is mentally doing with the information once they have access to it.
I could design a beautifully structured lesson that reduces unnecessary cognitive load, supports language and keeps everyone working.
Students could still process the content shallowly.
That possibility is important to me because it keeps CLSM from becoming a framework about making school easier. That has never been what I want it to become. Reducing unnecessary difficulty should create more room for the difficulty that matters.
The thinking.
And sometimes thinking looks unimpressive.
A student staring at a question for twenty seconds might be retrieving.
Someone crossing out an answer may have noticed a contradiction.
A student who asks, “Wait, wouldn't that mean...?” may be reorganizing an entire idea.
Another student may finish quickly because the task never required them to move beyond recognition.
If I only looked at productivity, I could easily misunderstand all four.
There is newer research that makes this even more interesting. A 2025 study examining observable student behavior argued that behaviors cannot always be assigned a fixed meaning because their relationship to engagement depends on the instructional context. The same outward behavior may tell us different things in different situations.
That seems almost painfully obvious once you say it, but schools make judgments from visible behavior constantly.
So do I.
We have to. A teacher cannot open thirty skulls and check what is happening inside. We use what we can observe.
The mistake would be believing that what we can observe is the learning itself.
I've been thinking about this in my own classroom because I intentionally build independent work time into lessons. There are periods when my classroom gets very quiet. Students may be working through guided questions, reading, using their Chromebooks or writing. There isn't necessarily a lot to watch.
Earlier in my career, I probably would have felt pressure to fill more of that silence.
Now I sometimes walk around and look for something different.
What are students actually doing with the information?
That question has changed what I notice.
A student going back three questions because something they just learned changed an earlier answer catches my attention. Someone checking their notes before committing to an explanation catches my attention. A student who stops me because two ideas seem to conflict catches my attention.
Those moments don't make a classroom look more engaging.
They give me better clues about cognition.
This also explains why retrieval has become such a large part of how I teach. The next day tells me things the previous day cannot.
I can leave class convinced that everyone understood something. Then I ask students to bring it back from memory and discover that what felt incredibly clear yesterday left only a faint trace.
I've had the opposite happen too. A lesson can feel ordinary while I'm teaching it, and then students retrieve far more than I expected later.
That has made me suspicious of my own perception of a “good lesson.”
The feeling in the room matters. Relationships matter. Interest matters. Participation matters. I want students curious. I want them involved. I would much rather teach a class where students care about what we're doing.
I just don't want to confuse the conditions around learning with evidence that learning occurred.
And this is where something happening in education right now makes the question even stranger.
Recent meta-analyses of generative AI report positive effects on student engagement, including cognitive engagement in many of the studies examined. At the same time, researchers continue to warn about over-reliance and emphasize that the way the technology is used matters enormously.
That creates a fascinating problem.
A tool can increase engagement while also making it easier to outsource parts of a task.
Technology can make students more involved without automatically telling us who is doing the thinking.
The same is true without technology.
A classroom can be energetic without being cognitively demanding.
It can also be quiet without being cognitively empty.
So I have started wondering whether we need to become much more precise when we say students are “engaged.”
Engaged with what?
The activity?
The other students?
The technology?
The reward?
The finished product?
Or the idea?
Those answers can overlap. I hope they often do.
But I don't think we should assume they do.
For me, the question has gradually changed. I still notice whether students are participating. I still care whether an activity catches their interest. I still pay attention to the atmosphere in the room.
I'm just much more interested now in what happens after the activity is over.
What can they explain when the slide disappears?
What relationship can they see that I didn't point out?
What do they remember tomorrow?
What can they do when I change the example?
What do they question?
That is where learning becomes harder to perform for us.
And maybe easier to actually see.
Originally published on LinkedIn on September 3, 2026.
One of the strangest things about teaching is that I can see almost everything a student is doing and still have very little idea what is happening in their mind.
I can see the Chromebook open. I can hear the discussion. I can watch someone highlight a reading, fill in notes, answer questions, move around the room, make a presentation and stay busy for an entire class period. If I walked past that student twenty years ago, I probably would have thought I was looking at engagement. I still might. What I am less certain about now is how much that tells me about learning.
That distinction has been bothering me lately because engagement has become one of those words in education that seems obvious until you try to define exactly what you mean by it.
Researchers don't necessarily treat engagement as one thing. Behavioral engagement can involve participation, attention and persistence. Emotional engagement gets closer to interest, enjoyment or a student's feelings about the learning. Cognitive engagement deals with something harder to observe: the mental effort being invested in understanding, connecting, reasoning, monitoring and making sense of information. Recent research continues to separate these dimensions because they do not always move together.
That makes classroom observation much more complicated than it first appears.
A student can be behaviorally engaged while doing relatively little with the content itself. Another student can look fairly passive while trying to connect a new idea to something learned yesterday. A conversation can involve deep reasoning or almost none. Highlighting can require a student to make decisions about importance, or it can become something the hand does while the mind is somewhere else.
The behavior doesn't tell the whole story.
There is research behind this problem that I wish I had understood earlier in my career. The ICAP framework, developed by Michelene Chi and Ruth Wylie, distinguishes passive, active, constructive and interactive forms of cognitive engagement. The interesting part is not simply that students should “do things.” The type of thinking elicited by the activity matters. Generating an explanation, making an inference or building on another person's reasoning asks something different of the mind than copying or manipulating information already provided.
Newer studies are still finding useful differences along those lines. A 2025 experiment involving instructional video found differences in cognitive load, retention and learning depending on the kind of interaction students experienced, with richer forms of interaction producing broader benefits.
I find that interesting because classrooms are full of activity now.
Students drag things. Click things. Create things. Discuss things. Respond to polls. Make slides. Watch videos with embedded questions. Collaborate on documents. Rotate through stations. Highlight digitally. Move information into boxes.
Some of those things can produce excellent learning.
Some can produce excellent-looking classrooms.
Those are not guaranteed to be the same classroom.
I have become much more interested in the cognitive work hiding underneath an activity. If students are making a presentation about an empire, for example, I want to know where the decisions are. Did I already tell them which information matters and where to put it? Are they mostly transferring information from one place into another? Or do they have to decide which developments mattered, connect political decisions to economic consequences, defend why one factor was more significant than another, or notice a relationship I never explicitly gave them?
The presentations might look almost identical when they're finished.
The thinking required to produce them could be completely different.
That gets very close to something I've been trying to understand through CLSM.
Cognitive load matters because students have limited capacity to work with information at one time. Language matters because students cannot reason deeply about information they cannot adequately access. Structure matters because organization can free attention for the content instead of making students spend it figuring out what to do. Memory matters because whatever happens during the lesson eventually has to survive beyond the lesson.
But none of those pieces tells me much if I never ask what the student is mentally doing with the information once they have access to it.
I could design a beautifully structured lesson that reduces unnecessary cognitive load, supports language and keeps everyone working.
Students could still process the content shallowly.
That possibility is important to me because it keeps CLSM from becoming a framework about making school easier. That has never been what I want it to become. Reducing unnecessary difficulty should create more room for the difficulty that matters.
The thinking.
And sometimes thinking looks unimpressive.
A student staring at a question for twenty seconds might be retrieving.
Someone crossing out an answer may have noticed a contradiction.
A student who asks, “Wait, wouldn't that mean...?” may be reorganizing an entire idea.
Another student may finish quickly because the task never required them to move beyond recognition.
If I only looked at productivity, I could easily misunderstand all four.
There is newer research that makes this even more interesting. A 2025 study examining observable student behavior argued that behaviors cannot always be assigned a fixed meaning because their relationship to engagement depends on the instructional context. The same outward behavior may tell us different things in different situations.
That seems almost painfully obvious once you say it, but schools make judgments from visible behavior constantly.
So do I.
We have to. A teacher cannot open thirty skulls and check what is happening inside. We use what we can observe.
The mistake would be believing that what we can observe is the learning itself.
I've been thinking about this in my own classroom because I intentionally build independent work time into lessons. There are periods when my classroom gets very quiet. Students may be working through guided questions, reading, using their Chromebooks or writing. There isn't necessarily a lot to watch.
Earlier in my career, I probably would have felt pressure to fill more of that silence.
Now I sometimes walk around and look for something different.
What are students actually doing with the information?
That question has changed what I notice.
A student going back three questions because something they just learned changed an earlier answer catches my attention. Someone checking their notes before committing to an explanation catches my attention. A student who stops me because two ideas seem to conflict catches my attention.
Those moments don't make a classroom look more engaging.
They give me better clues about cognition.
This also explains why retrieval has become such a large part of how I teach. The next day tells me things the previous day cannot.
I can leave class convinced that everyone understood something. Then I ask students to bring it back from memory and discover that what felt incredibly clear yesterday left only a faint trace.
I've had the opposite happen too. A lesson can feel ordinary while I'm teaching it, and then students retrieve far more than I expected later.
That has made me suspicious of my own perception of a “good lesson.”
The feeling in the room matters. Relationships matter. Interest matters. Participation matters. I want students curious. I want them involved. I would much rather teach a class where students care about what we're doing.
I just don't want to confuse the conditions around learning with evidence that learning occurred.
And this is where something happening in education right now makes the question even stranger.
Recent meta-analyses of generative AI report positive effects on student engagement, including cognitive engagement in many of the studies examined. At the same time, researchers continue to warn about over-reliance and emphasize that the way the technology is used matters enormously.
That creates a fascinating problem.
A tool can increase engagement while also making it easier to outsource parts of a task.
Technology can make students more involved without automatically telling us who is doing the thinking.
The same is true without technology.
A classroom can be energetic without being cognitively demanding.
It can also be quiet without being cognitively empty.
So I have started wondering whether we need to become much more precise when we say students are “engaged.”
Engaged with what?
The activity?
The other students?
The technology?
The reward?
The finished product?
Or the idea?
Those answers can overlap. I hope they often do.
But I don't think we should assume they do.
For me, the question has gradually changed. I still notice whether students are participating. I still care whether an activity catches their interest. I still pay attention to the atmosphere in the room.
I'm just much more interested now in what happens after the activity is over.
What can they explain when the slide disappears?
What relationship can they see that I didn't point out?
What do they remember tomorrow?
What can they do when I change the example?
What do they question?
That is where learning becomes harder to perform for us.
And maybe easier to actually see.
Originally published on LinkedIn on September 3, 2026.