It’s 1:04am. The exam is at 9. You’ve been lying here since 11:30. You know you should be asleep. You’re doing the math — if you fall asleep right now you’d get seven and a half hours, except you’re not falling asleep right now, so now it’s seven hours and forty-five minutes and counting, except you’re still not asleep, and the thing you’re aware of most acutely is that you are not asleep.
You’ve tried the breathing. You’ve tried flipping the pillow. You’ve told yourself to stop thinking about it. You are still thinking about it.
Here’s what’s actually happening, and what to do about it — which is not “try harder to sleep.”
The effort paradox
Sleep is one of the few things that gets harder when you try. You can’t force yourself to sleep the way you can force yourself to finish a problem set. Falling asleep is an automatic process — it happens when the conditions are right and you’re not actively interfering with it. The moment it becomes a project you’re managing, you’ve already disrupted it.
Sleep researcher Colin Espie describes this as the attention-intention-effort pathway — a mechanism by which normal sleepers become poor ones. The model goes like this: you have a bad night, so you start paying more attention to whether you’re sleeping. Paying attention creates an intention to sleep. Intending to sleep creates effort. Effort increases cognitive and physiological arousal. Arousal prevents sleep. And now you’re in a loop that reinforces itself.
The exam-night version of this is compressed and acute. You’ve got a reason to care urgently about sleep tonight, so you’re monitoring it closely, which activates the exact mechanism that keeps you awake.
To make it worse: you probably had caffeine after 4pm, which has a half-life of about five to six hours — meaning half the caffeine in that 6pm coffee is still circulating at midnight. You probably spent the last hour or two studying on a bright screen. You’re physically still, which means the adrenaline that showed up for the exam is sitting in your body with nowhere to go. All three of these are genuine physiological factors, not just psychology.
The solution almost everyone tries first is to lie there and keep attempting sleep. This is the worst option. It extends the time you’re lying in bed awake, which trains your brain to associate your bed with wakefulness and effort rather than with sleep — the core mechanism behind chronic insomnia. You’re not at risk of chronic insomnia from one bad night, but you are making the current night worse with every additional hour of lying there trying.
What Harvey’s model tells us about pre-sleep arousal
Allison Harvey’s 2002 cognitive model of insomnia maps the sequence precisely. Excessive cognitive activity before sleep — worry, planning, replaying — produces autonomic arousal and emotional distress. This creates a feedback loop: the arousal makes it harder to sleep, which produces more worry about not sleeping, which sustains the arousal.
The pre-exam brain is running exactly this process. You’re not consciously worrying about everything — maybe you’re not thinking about the exam at all anymore. But the arousal is already present, and your attempt to monitor your sleep is maintaining it. The cognitive system is in input mode — alert, scanning, processing — and input mode is incompatible with sleep onset.
The model also describes something called “safety behaviors” — things people do to try to control their sleep that paradoxically maintain the problem. Lying still and waiting for sleep is a safety behavior. Checking the clock is a safety behavior. Trying to calculate exactly how much sleep you’ll get if you fall asleep right now is a safety behavior. Every one of these keeps your brain in active monitoring mode.
What to actually do at 1:15am
There’s something worth saying about the exam itself before you do anything else. A single bad night’s sleep before an exam is far less damaging than most students believe. Decades of research on acute sleep deprivation show that one night of poor sleep significantly impairs sustained attention and processing speed — but it impairs them measurably less than the anxiety response you generate by lying awake trying to force sleep and catastrophizing about being tired. The most prepared students who sleep five hours and wake up calm tend to outperform equally prepared students who sleep six hours and arrive at the exam hall still wound up from an hour of lying-awake-effort the night before. The sleep matters. The panic about the sleep matters too.
This doesn’t mean you should stop caring about sleep. It means you should stop making the active effort to sleep be a second exam you have to pass tonight.
The most counterintuitive and most evidence-consistent thing you can do: get up.
Not to study. Not to check your phone. Get up, go to a different room — the common room, the hallway, the bathroom if that’s the only option — and find something boring to read on paper. A physical textbook from a different course. A novel you don’t care about. A magazine. Anything that keeps your eyes moving without triggering the alert, connected feeling that screens produce.
Read for fifteen to twenty minutes. Don’t check the time. Don’t think about the exam. When you feel genuinely drowsy — not just tired, but heavy-eyed and slow — go back to bed.
The mechanism behind this is called stimulus control, and it’s one of the most reliably effective behavioral interventions for sleep difficulty. The goal is to break the association between your bed and lying-awake-trying. You go back to bed when sleep is actually likely, not while you’re in an active effort loop.
A few specifics for exam night:
Don’t study. This feels counterproductive but isn’t. The marginal value of thirty more minutes of studying at 1am is functionally zero — your memory consolidation is compromised when you’re in this state, and you’ll recall it worse tomorrow than if you’d slept. You’re not going to learn something crucial in the next hour that saves you.
Leave your phone in the room. The common room trip only works if you’re not carrying a glowing screen with you. The light exposure and the stimulation both reset the arousal upward.
When you return to bed, give up the project. Don’t try to fall asleep. Lie there with the explicit intention of resting rather than sleeping. Often sleep follows from this. Sometimes it doesn’t. If you’re still awake thirty minutes later, repeat the process — get up, different room, boring paper reading, return when drowsy.
The honest version of this: you may not get eight hours tonight. You might get five. Five hours the night before an exam, combined with low anxiety about the sleep itself, will likely perform better for you than eight hours spent half of them lying in bed in an effort spiral. Sleep deprivation affects you, but acute anxiety and hyperarousal affect your recall and processing too. The effort loop is part of the problem.
One more thing. The most consistent predictor of a bad exam-night is what happened the two nights before. A single bad night on top of two normal nights is manageable. A bad night on top of several nights of sleep debt is harder. If you’re pattern-matching to the second situation, recovering from an all-nighter covers what to do in the days after sleep debt accumulates.
For tonight: 1:15am, get up, different room, boring book, lights low, no phone. Come back when you actually feel sleepy. Let the rest happen without effort.