Attention and divided focus

Also known as multitasking, media multitasking, distraction

Nothing you did not attend to gets learned, and attention does not split — it alternates, expensively.

The claim Learning requires attention at the moment of encoding; dividing attention between tasks degrades what is learned, and the cost is larger than people estimate.

Good evidence That divided attention harms encoding is not seriously disputed; several popular specifics — notably the smartphone 'brain drain' effect — have failed to replicate.

Attention is the gate. Material that does not pass through it is not weakly learned, it is not learned at all, and this is the least negotiable constraint in this whole section. Every technique on the other pages assumes you were paying attention when the material went in.

What people call multitasking is, for anything that demands thought, rapid alternation with a switching cost paid each time. The cost is not merely time. Divided attention during encoding produces memories that are shallower and less well connected, and — the part that surprises people — it does so without much affecting how well you think you learned, which is the illusions-of-learning problem again in a different costume.

The most useful classroom evidence is Sana, Weston and Cepeda's 2013 study, for a reason beyond its main finding. Students multitasking on a laptop during a lecture scored worse on a test afterwards, which is unsurprising. Students who were merely in view of someone else's multitasking laptop also scored worse — a cost they did not choose, could not see and would not have reported.

The single most-quoted finding in this area deserves less confidence than it gets. Ward and colleagues reported in 2017 that a smartphone lying face-down on the desk reduced working-memory capacity even when never touched — the 'brain drain' effect, repeated everywhere since. A preregistered direct replication in 2022 found nothing: phone placement made no difference to working memory or inhibitory control. The honest summary is that the specific claim is unresolved, while the broader finding that actually using a phone mid-task hurts learning is not in doubt.

This is also where cognitive load theory and attention meet. Working memory is the resource both are describing; a distraction and a badly designed exercise consume the same limited thing. The practical implication is the same in both literatures: protect capacity for the material, and spend nothing on anything else.

What to do differently

  1. Do the encoding — the first real contact with new material — in a single block with nothing else running. Review can survive interruption; first learning cannot.
  2. Close the other tabs rather than resolving to ignore them. Suppressing an urge uses the capacity you were protecting.
  3. Prefer short undivided sessions to long divided ones. Twenty-five focused minutes beat ninety interrupted ones, and are easier to actually do.
  4. Treat background media as either study or entertainment, never both — a podcast you are half-listening to is doing neither job.
  5. Judge focus by output, not by how long you sat there. Time in the chair is the least informative measure available.

Limits and trade-offs

  • The famous smartphone-presence result failed a preregistered replication, and anyone citing it as settled — including advice sites — is overstating the evidence.
  • Most laboratory work uses short tasks and young adults, and the size of the multitasking penalty for long self-directed study is not well measured.
  • Attention research describes a cost; it does not tell you how to have more attention, and most of what is written about that is untested.
  • Undivided study is a privilege of circumstances. Advice that assumes a quiet hour is useless to someone learning in twenty-minute fragments — and fragmented study still works far better than none.

What this means for a language

The specific trap here is passive listening. Playing a podcast in a language you barely speak while you cook is close to worthless for acquisition, because the input never gets attended to and unattended speech does not become comprehensible input — it becomes noise you have grown used to. That is a real distinction and not a purist one: the same podcast, listened to properly for fifteen minutes, does more than three hours in the background. The exception is a narrow one, and worth keeping: pure sound exposure may help you get used to the rhythm and phonology of a language, which is a different and much smaller claim than learning it.

Methods that lean on this Narrow Listening Comprehensible Input

The research

Related

Builds on Worked examples and cognitive load Works with Illusions of learning Works with Sleep and memory

Read next

The full reading list →

← All learning principles