In 1927, at Barnard College in New York, a young psychologist named Arthur Jersild handed seventy students sheets of two-digit numbers and started a laboratory stop clock.
In one column, add 6 to every number. In another, subtract 3. In a third, alternate: add, subtract, add, subtract.
The pure columns took about 85 seconds on average. The alternating ones took about 115. A third more time for exactly the same arithmetic, only with the rule changing on every line.
So far, nothing you would not have guessed. The interesting result came next. When Jersild alternated subtracting 3 from a number with giving the opposite of a word, the alternating column came out faster than the pure ones. You cannot give the opposite of a number or subtract 3 from a word: each item announced which task it wanted.
Nearly a century later, that detail is still the most useful thing known about task batching. Almost nobody mentions it.
Three names for one question
Multitasking, task switching and batching sound like three techniques. They are three answers to the same question: what to do about the switch.
What each one really is
Multitasking, when both things need attention, is a generous name for rapid switching. The exception is real but narrow: if one of the tasks is automatic, like walking or folding laundry, you can do them together. When both require thought, there is no parallel processing, only taking turns.
Task switching is what psychology has been measuring since Jersild. Joshua Rubinstein, David Meyer and Jeffrey Evans, in four experiments published in 2001, described a switch in two stages: shifting the goal ("I want to do this now") and activating the new rules ("off with those, on with these"). Each activation cost tenths of a second, and the cost grew when the tasks were more complex or less familiar.
Batching is not a psychology concept. The word comes from batch production in manufacturing and reached personal productivity as advice: replies in one slot, calls in another. The logic is sound. The direct evidence, as you will see, is thin.
What the research supports, and what it does not
Switching costs even when you prepare. Stephen Monsell's 2003 review pulls together decades of experiments: knowing in advance that a switch is coming reduces the cost, but does not remove it. A residual cost remains that preparation cannot reach. Switching less often is therefore the only way to pay less.
The cost lives in rule confusion, not in change itself. That is Jersild's message. Switching between things that cannot be confused is almost free; switching between similar things with different rules is what hurts. That changes what a good batch is. Ten message replies make a batch because the rule is the same each time. A "project X block" with email, a spreadsheet, a text review and a phone call has one topic and five different rules. It looks batched and behaves like switching.
Switching mid-task leaves a trace. Sophie Leroy showed in 2009 that people who leave a task unfinished to start another carry attention residue: part of their mind stays on the previous task, and performance on the next one drops. A batch only saves something if it ends at a clear stopping point. A batch broken off halfway is switching with extra steps.
Batching your checks brings relief, not proven output. Kostadin Kushlev and Elizabeth Dunn asked 124 adults to check email only three times a day for one week and as often as they liked in another. In the limited week, daily stress was lower. But there was no direct effect on productivity, and on average participants could not stick to the rule exactly.
Which one to use when
| Multitasking | Task switching | Task batching | |
|---|---|---|---|
| What happens | Disguised turn-taking | Rule change, with a cost | Fewer, planned switches |
| When it fits | One task is automatic | Very different tasks | Repetitive tasks with the same rule |
| Main cost | Errors and fatigue | Reconfiguration time | People waiting on your reply |
| Evidence | Against, for tasks that need attention | Strong, in the lab | Indirect |
The limits of what we know
I found no controlled study testing task batching as a personal productivity technique. What exists is lab research on switch costs, using tasks that last seconds, and one email experiment that measured stress, not output. The bridge between them is reasonable, but it is a bridge, not a finding.
And batching has a cost of its own: whoever is waiting for your answer waits longer. In roles where replying fast is the job, batches need to be smaller.
Where this meets ASB
What tripped me up was never a lack of batches on my calendar. It was the idea that arrived in the middle of the wrong batch and made me switch tasks just so I would not lose it.
That is why ASB lives in WhatsApp: you send the thing in one sentence the moment it shows up, and go back to what you were doing. It holds it and hands it back when the right batch comes round.