You are twenty minutes into a task that will take forty. The work is steady, the next real decision is still some distance off, and somewhere between one keystroke and the next you are thinking about a conversation from Tuesday, or a bill, or nothing you could name afterward. Then the task tightens — the part that actually requires you — and you are back, without quite noticing you left or that you returned.
Most of the language we have for that moment is unkind. Distracted. Not focused. Somewhere else. But a preregistered study of 269 adults published in Psychonomic Bulletin & Review suggests something less like a lapse and more like an allocation. Matthew Welhaf, Julie Bugg, and Jonathan Banks set out to test whether people time their mind wandering — spending it when a task is cheap and reclaiming it as the task gets expensive — and whether some people are better at that timing than others. They found that the timing is real, and that a specific cognitive ability predicted who did it most sharply. Not fluid intelligence. Not how much someone says they mind-wander in daily life. Working memory capacity.
A rotating clock hand and the question "where were you just now?"
The task the researchers used is deliberately plain. A clock hand rotates around a face, and participants are asked to respond when it reaches the 12 o’clock position. Nothing about the first three-quarters of the rotation requires much. The demand is concentrated at one predictable point, and it arrives on schedule, over and over.
Scattered through the task are thought probes — interruptions that ask participants where their attention actually was in the seconds just before. That is the measurement problem mind-wandering research keeps running into: attention that has drifted rarely announces itself, and asking about it afterward mostly retrieves what people believe about themselves. Probing in the moment is imperfect too, but it catches the state closer to when it happened, and it can be repeated enough times to build a pattern rather than an impression.
The pattern is what makes the design work. Because the probes can be sorted by where the clock hand was — how near the rotation had come to the demanding 12 o’clock moment — the researchers can look at whether reported mind wandering shifts across the cycle. Strategic modulation, in this design, has a specific shape: more mind wandering while the hand is still far from the target, and less of it as the hand closes in. That is the empirical fingerprint of budgeting. Attention spent where it is not needed, then pulled back before the bill comes due.
This idea did not originate here. The study is a preregistered replication and extension of earlier work by Seli and colleagues from 2018, which reported that people modulate mind wandering with task demand. Preregistration matters for a claim like that: the analysis plan is committed to in advance, which makes it much harder for a pattern to emerge from the flexibility of the analysis rather than the data.
The replication held — and only one individual difference tracked it
The strategic pattern replicated. Across the sample, mind wandering was more common in the low-demand stretches of the rotation and receded as demand approached its peak. That is the part that confirms rather than surprises, and it is worth saying plainly that confirmation is the point of a preregistered replication, not a consolation prize.
The extension is where the new information sits. The researchers asked which individual differences predicted the size of that shift — who modulated a lot, who modulated a little. They tested working memory capacity, fluid intelligence, and trait mind wandering measured through dispositional scales that separate spontaneous from deliberate drifting.
Only working memory capacity predicted it. Participants with higher WMC showed a larger swing away from mind wandering in the moments before the clock task’s demands peaked. Fluid intelligence did not track the pattern. Neither did how much someone reported mind-wandering as a general tendency, in either the spontaneous or deliberate flavor.
That last null is quietly interesting. The dispositional measures ask people to characterize their own attention, and the strategic pattern is something people do in seconds without necessarily narrating it. A person can believe they drift constantly and still time it well. A person can believe they are focused and still not modulate much. Self-report about a habit and moment-to-moment control of that habit appear, here, to be different things.
What a clock task can and cannot tell us
This is a laboratory task with an unusually generous structure. The demand is predictable, periodic, and short. Nothing about it resembles the ragged demand curve of a workday, a conversation, or a drive — where the expensive moment can arrive without a rotating hand to announce it. The study does not show that people budget attention well in those settings, and it does not measure whether budgeting there helps or costs anything.
What it does support is narrower and still worth holding onto. In a setting where the timing of demand is learnable, mind wandering behaved less like a control failure and more like a resource being spent and recalled — and the people who managed that spending most sharply were the ones with more working memory capacity to manage it with. The authors frame it as evidence that mind wandering is not necessarily a breakdown of control; it can be under control, particularly for some people.
Which leaves the reframe rather than the tip. The drifting itself may not be the diagnostic thing about attention. When it happens, and whether it ends before it costs you, might be. That is a less flattering story for anyone who wants a rule to follow, and a kinder one for anyone who has spent years reading their own wandering as a character flaw. Some of it, at least in a room with a rotating clock hand, looks like arithmetic.
