Jet lag: the hidden cost of switching contexts

Every move between a subject, an app, a chat window and a task costs you altitude. The cost stays invisible because it is not paid during the move. It is paid in the first minutes afterwards, when part of your head is still with the thing you just left. The switch feels free. The bill arrives later, with nothing on it that says what it was for.

Why juggling feels efficient

Nothing in a study session runs in parallel. Reading a proof and answering a message take turns in slices, and every slice has a seam at each end. What feels like doing two things at once is a fast sequence of small departures and landings.

It feels efficient because the small tasks close. A message gets answered, and you can see that it is answered. A statistics chapter does not close in thirty seconds, and twenty minutes of work on it often leaves nothing visible. So a session with eleven interruptions produces eleven small completions and one unfinished chapter, and the feeling afterwards is that a lot got done.

There is a less flattering half to this. Switching is most tempting exactly where the material gets hard. Leaving the paragraph you do not understand looks like efficiency from outside and is usually relief from inside.

The arithmetic nobody does

People defending a quick look at their phone count the wrong quantity. They count the thirty seconds spent inside the chat, and that part is genuinely cheap. What they leave out is the stretch between putting the phone down and the moment the sentence in front of them makes sense again.

It is worth measuring once. Note the time you return to the page, and the time you are following the argument again rather than reading past it. The second number is usually larger and bears no relation to the length of the interruption. A ten second glance and a two minute reply cost about the same on the way back, because the expensive part was never the detour. It is the reconstruction afterwards.

Attention residue: part of you is still back there

What makes the return expensive has a plain name. After a switch, part of your attention stays with the previous task. You are sitting at the statistics problem, but something in the background is still composing the reply you never finished typing. Research on task switching generally points this way: leaving a task does not clear it out of your head, and it clears out even less if the task was left open.

This is why an interruption arriving as a question is worse than one arriving as information. A message saying the seminar was moved is closed the moment it is read. A message asking whether you can take over the presentation stays open and keeps drawing a small share of attention until it is decided.

The practical consequence is not that you should never look at anything. Open loops belong on paper rather than in your head. One written line next to the keyboard, and the loop stops running in the background.

One subject per flight

A block labelled study that contains statistics, the seminar paper and an email to the examination office is three sessions sold as one. Each has its own vocabulary, its own way of thinking and its own state to rebuild. Moving between them inside ninety minutes means paying the entry cost three times and finishing none of them.

One subject per session is a crude rule and it holds up well. What matters is the moment a second subject shows up uninvited, which it will. Halfway through a derivation you remember that the literature list is due on Friday. The move that keeps the session intact is to write the words literature list on a sheet of paper and go back. It works because the thought does not need to be solved now. It only needs to stop being in danger of getting lost.

Batch the admin instead of sprinkling it

Administrative tasks are individually short and collectively expensive, and they are the most common source of self inflicted switches. Uploading a scan to the course platform, replying to a tutor, booking a library seat, checking whether the room changed: none of it takes long, and each item lands in the middle of something if allowed to.

The fix is unremarkable. Collect them somewhere during the day and do them in a single slot, ideally at an hour when you would not be thinking well anyway. A day with one admin slot and two clean blocks gets more done than the same day with the same tasks spread evenly across it.

A switch at a seam is not the same as a switch mid thought

Two sessions can contain the same number of switches and cost completely different amounts. What separates them is where the switch lands. At a seam, when a problem is solved or a chapter is finished, nothing is left in working memory that has to be held. Standing up, answering two messages and coming back is close to free.

In the middle of a derivation the situation is the opposite. You are holding four or five things that exist nowhere except in your head: which substitution you chose, why the second case behaves differently, what you were about to check. None of that survives the trip to the phone. You come back to the paper, not to the state you were in when you left it.

So the useful discipline is not fewer switches in the abstract. It is deciding when they happen. A planned switch at a seam is a landing. An unplanned one in the middle of a thought is an interruption with a nicer name.

The switch that is not a switch

The advice above gets misapplied in one particular way that deserves naming. Deliberately mixing different types of task inside one subject is not context switching at all, and while you are practising it is often better than working in blocks.

Alternating between three kinds of problem in a mixed set, moving between a derivation and a worked example, going from vocabulary to grammar inside the same language: in all of these the context stays put. The notation, the concepts and the state in your head carry straight across. What changes is which method you have to select, and having to select it is precisely the skill an exam tests. Work on practice schedules points the same way: mixed practice feels worse while it happens and tends to hold up better afterwards, while thirty problems of one type in a row feel smooth partly because the heading already told you what to do.

So the distinction is not switching against not switching. It is whether the switch changes the context or only the demand inside one context, and whether you chose the moment or something else chose it for you.

The second limit is practical. For a student with a child or a working student on shift, the honest answer is not to switch less, because switching less is not on offer. It is to make the return cheap: stop at a visible seam and leave a written line about the next concrete action.

The part that actually changes things

None of this needs a system. It needs the cost counted in the right place. Once the expensive part of an interruption is understood to happen after the interruption, leaving the phone in the other room stops looking like an overreaction.

A session that keeps to one subject, ends its switches at seams and pushes admin into a single slot is not more disciplined than the alternative. It is simply cheaper to fly.