THE REAL CURRICULUM

The Real Curriculum Is What You Model Under Stress

By Seçil Sayhan, MSc Clinical Health Psychology & WellbeingUpdated July 2026

The takeaway

the lesson isn't what you say under stress, it's what you do while everyone's watching

What’s in this article

  1. You teach with your worst ninety seconds
  2. Why stress turns you into the textbook
  3. Why "I'll explain it later" doesn't work
  4. Narrate your steps out loud
  5. This isn't about never breaking
  6. The frontier is your own regulation
  7. Frequently asked questions

In 1961, Albert Bandura sat children in front of an adult who attacked an inflatable doll, then left them alone in the room. No instructions. No rewards. The kids who watched the aggression repeated it, sometimes gesture for gesture. The finding has held for sixty years, and the part nobody tells you is that it doesn't stop in childhood. We copy hardest in the moments things fall apart.

You teach with your worst ninety seconds

We like to believe we teach the people around us with our words. The values we state out loud. The advice we hand over when someone's stuck. The quote we share when a friend is having a hard week.

We mostly teach with our worst ninety seconds.

Think about who actually shaped how you handle pressure. It probably wasn't a speech anyone gave you. It was the way your father went quiet and cold when money got tight, or the way a first boss stayed level while a launch fell apart and the room kept its head because hers stayed on. You weren't being taught. You were watching. And watching was enough to install it.

This is the gap that trips up thoughtful people. They invest enormous care in the stated curriculum: the talk about resilience, the speech about staying calm, the framed value on the wall. Meanwhile the real curriculum runs underneath it, on a separate channel, and it's made entirely of what you do when you didn't plan to be observed. The bad email. The number that came back wrong. The plan that died on a Tuesday. Whatever you did in that window is the lesson that landed.

Why stress turns you into the textbook

Here's the mechanism, and it's physiological, not moral.

Stress narrows attention. When a room gets tense, every nervous system in it does the same ancient thing: it scans for the most senior, most steady person present and treats that person as the reference signal. Are we safe or not? The answer doesn't come from what you say. It comes from your face, your shoulders, how fast your voice climbs, how quickly you reach for blame. The body reads the body first, and it reads it faster than language.

There's a second layer underneath that. Our brains are wired to mirror the people we watch, especially people we rank as competent or in charge. We rehearse their actions internally just by observing them. Under calm conditions that mirroring is gentle background noise. Under threat it sharpens, because a stressed brain is hunting for a template to copy so it doesn't have to compute a response from scratch.

So the worst possible moment to be careless about your behavior is the exact moment your behavior gets recorded most deeply. That's not a coincidence. That's the design. The people watching you aren't grading your intentions. They're downloading your regulation, or your lack of it, straight into their own.

Why "I'll explain it later" doesn't work

The standard repair move is the after-action talk. You lose it in the meeting, then later you sit the person down: that's not how I want to handle things, here's what I should have done, here's what I actually value. Honest. Well-meant. Mostly useless as a teaching tool.

It fails because the two messages don't carry equal weight. The blow-up was absorbed in real time, with the nervous system wide open, attention narrowed to a point. The explanation arrives later, in calm, through language, to a person who has already filed the lesson. You're trying to overwrite a felt experience with a verbal correction. The felt experience wins almost every time.

This is also why "do as I say, not as I do" never produced a single calm person. Kids don't learn pressure-handling from instruction, and neither do adults. A team will quote your values back to you in the all-hands and still flinch the way you flinch when the quarter goes bad, because flinching is what they actually saw.

The explanation isn't worthless. It does real work after the modeled behavior was already sound, as a way to name what you did. But used as a patch over the thing they watched, it's a receipt for a purchase nobody made.

Narrate your steps out loud

There's a small, usable move here, and it costs nothing.

The next time something goes sideways in front of people, narrate your first steps out loud. Plainly. "Okay, this is a real problem. Here's what I'm doing first: I'm going to find out what actually broke before I decide anything." Then do that.

You are not performing composure. Performed calm reads as fake and teaches people to hide their stress instead of working through it. What you're doing is making your process visible. You're handing the room a template it can copy: name the problem, slow the first move, act on information instead of adrenaline. That sequence is the thing worth installing, and saying it out loud is how it gets installed.

Three concrete versions. The number comes back wrong: "I don't like this number. First thing, I'm checking whether it's the data or the reality." The plan collapses an hour before it ships: "We're not making the deadline the way we planned. I'm deciding what's the smallest version we can stand behind." Someone sends the message at the worst possible time: a beat of silence before you answer, visible, so the room learns the beat exists.

The goal isn't to look unbothered. It's to let people watch a stressed person stay functional, because that's the only version of this skill they can actually use.

This isn't about never breaking

Here's the objection, and it's a fair one. Doesn't this just load more pressure onto the person already under pressure? Now I can't even have a bad moment without warping someone?

No. The point isn't to never break. You will. Everyone does, and a leader or parent who is visibly never rattled teaches a quieter, worse lesson: that staying human under load is something to suppress and conceal.

The repair is part of the curriculum too. When you do snap, the recovery is watched as closely as the snap. "I came in hot just now, that was me being stressed, not you, and here's what I actually want to do." That models something most people never see done: a person noticing their own reaction, owning it without drowning in it, and re-entering the problem. That's arguably more useful than never reacting, because it's the realistic version. Nobody gets the unbroken-calm life. Everybody gets the chance to recover well.

What you can't do is be careless in the heat and silent in the aftermath. The breakage without the repair is the full lesson, and it's the wrong one. The breakage plus a clean recovery is a different lesson entirely, and it's one of the most valuable things you can pass on.

The frontier is your own regulation

Pull back and the implication is bigger than parenting or management. Your nervous system is a broadcast. The people around you are tuned to it, more than to anything you say, and they're tuned in hardest exactly when you'd most like them not to be looking.

This is why working on your own regulation is not self-indulgence. It's the most leveraged thing you can do for everyone downstream of you, because they inherit your defaults. The calm you build privately becomes the calm they have access to publicly. The reactivity you never address becomes the reactivity they reproduce, often in people who never met you, a generation or a hierarchy removed.

It also reframes what "setting an example" means. It was never the example you set on purpose, in the speech, on the good day. It's the example that escapes you on the worst day, and the only way to change it is to change what's actually in you to escape. You can't fake your way past your own physiology under load.

MARSA's NextGen program is built on exactly this idea: that the most durable thing you ever pass on is your regulated nervous system, and that it's trainable. If that's the work you want to do, it's at marsa.ai. But the move itself you can start today, for free, the next time something breaks while people are watching.

The most powerful thing you teach is what you do in the moment you forgot you were being watched.
we built NextGen for exactly this, raising steady humans in an unsteady world, the science of what gets modeled and how to model it on purpose ($247). deeper read on the blog if you want the full breakdown of Bandura's work first.
Explore NextGen →

Frequently asked questions

What was the Bandura Bobo doll experiment, in plain terms?

In 1961, Albert Bandura had children watch an adult interact with an inflatable doll, either aggressively or calmly, then left each child alone in the room. The kids who'd watched aggression reproduced it, often copying specific gestures, with no instruction and no reward offered. It became the foundation of social learning theory and showed that we acquire behavior just by observing it, not only by being taught or rewarded.

Does observational learning really continue into adulthood?

Yes. Bandura's work started with children, but the underlying mechanism, our tendency to mirror the people we watch, especially those we see as competent or in charge, doesn't switch off with age. If anything it gets more selective: adults copy the steady, senior person in a tense room. The copying intensifies under stress, when a pressured brain looks for an existing template to follow instead of building a response from scratch.

Why does stress make people copy behavior more strongly?

Stress narrows attention and orients the nervous system toward whoever it reads as the reference point for safety. In that state the brain is actively hunting for a behavior to model so it doesn't have to compute one under load. That's why your reaction in a crisis lands deeper than any calm explanation later: it's absorbed in real time, with attention fully open, by people looking to you to tell them whether things are okay.

Isn't faking calm under pressure dishonest and obvious?

It is, and it teaches the wrong thing. Performed composure reads as fake and quietly trains people to hide their stress rather than work through it. The alternative isn't acting unbothered. It's making your process visible: name the problem out loud, slow your first move, act on information instead of adrenaline. You're letting people watch a stressed person stay functional, which is the only version of the skill they can actually use.

What if I lose my temper anyway?

You will, and that's not the failure. The recovery is watched as closely as the reaction. Saying "I came in hot, that was my stress, not you, here's what I actually want to do" models something most people never see: noticing your own reaction, owning it without drowning in it, and re-entering the problem. The thing to avoid is being careless in the heat and silent afterward. Breakage plus a clean repair is a genuinely valuable lesson.

How is this different from just teaching people good values?

Stated values run on one channel; your behavior under stress runs on another, and the second one carries more weight. People will quote your values back to you and still flinch the way you flinch when things go bad, because flinching is what they actually watched. You can't out-talk what you model. The only real fix is changing what's in you to escape under pressure, which is why regulation work matters more than any speech.

About the author

Seçil Sayhan is a behavioral scientist and the founder of MARSA.AI. Trained on both sides of her field — a BA in Business Management, an MSc in Clinical Health Psychology & Wellbeing, a diploma in neuroplasticity, and advanced training in Lifestyle Medicine from Harvard University — she has spent the past decade helping 7,000+ people across 12 countries rewire the systems running their lives. Behavior is one science — whether it moves a person, a market, or a machine. See the full bibliography at marsa.ai/research.