The dishes are done and the kitchen is quiet and the thing you have not told anyone arrives, uninvited, the way it does. Nobody asked. There is no conversation to survive, no face to hold steady. You are alone with the tap running and there it is again.
Most of what has been written about secrets is about the other moment: the interrogation you have to get through, the question you saw coming, the effort of keeping your expression level. That moment is real. It is also, according to the research that has actually tried to count both, the less frequent one.
What the researchers actually counted
In 2017, Michael Slepian, Jinseok Chun and Malia Mason published ten studies in the Journal of Personality and Social Psychology built on a change of definition. Earlier work had treated a secret as an act: something you do in a conversation, by withholding. They treated it instead as an intention. You have a secret from the moment you decide to hold something back, whether or not anyone ever asks.
That change makes a different question askable. If a secret exists in the gaps as well as in the conversation, how much of each is there?
To find out they first needed to know what people actually keep secret, so they asked 2,000 people to describe a current secret and sorted the answers into 38 categories: theft, mental health, an ambition, a surprise, a habit or addiction, finances. Then they handed that list to new samples of 200 and asked, per category, whether the person had ever had the experience and whether it was secret. Across three such samples, 96, 96 and 97.5 percent were currently keeping at least one. On average people were holding 13 of the 38, five of which they had never told anyone.
Then came the counting. Eight of the ten studies counted both. In their secrecy conditions, participants were asked for the same pair of numbers: the times they were with the person they were hiding it from and had to actively withhold it, and the times they were not with that person and found themselves thinking about it anyway. Most asked people to estimate across the past 30 days; one, described below, ran the count forward instead, and the largest, of 600 people, added the other two cells of the grid, thoughts that arrived while with the person and tracks covered in their absence. The ordering held across that grid too, though only barely in the person’s presence, and thinking of the secret while with the person also went with feeling worse; hiding it did not.
The second number was consistently larger. Modelling both across roughly 2,600 secrets per study, the researchers estimated that in a month people concealed a given secret 1.77 times and thought about it unprompted 4.24 times; in the replication, 2.44 and 4.82. Counted the way count data ought to be counted, that came out at 2.44 and 1.84 times more thinking than hiding. Of all the categories across both studies, the authors report that every one but surprises was thought about more than it was concealed.
The hardest case they could find
An obvious objection is that this is rigged by distance. If your secret is being kept from a cousin two states away, of course you rarely have to hide it in conversation. You rarely have a conversation.
So the same team ran the test on people in committed relationships, reporting on something they were keeping from their partner. In one such study, asked for a secret they felt guilty about, 128 people estimated thinking about it away from their partner an average of 13.05 times in the past month, against 7.00 times having to conceal it while talking to them. A second study, which asked instead for a secret that was simply significant, got 14.87 against 5.89.
And because a month is a long time to remember accurately, they ran it forward instead of backward, asking people to report both counts each night. A hundred started; after exclusions eighty went on to the nightly reports, and fifty-one lasted all five days. Across those nights the ordering held at the daily scale too, at a modelled 0.92 concealments a day against 2.09 unbidden thoughts, which the authors summarise as thinking about a secret twice as often as working to hide it.
The well-being half of the finding is the part worth stating carefully. When the two frequencies were entered together as predictors, how often a secret came to mind unbidden was related to people saying it had made their life worse; how often they had to hide it was not. The concealment coefficient sat near zero every time, and in the significant-secret study it predicted neither satisfaction with life nor how the relationship was going, while the thinking predicted both.
Note what that does not say. It does not say concealment is pleasant. In that same study, participants rated the moments of hiding as carrying less positive feeling than the moments of unbidden thinking, though the two were indistinguishable on negative feeling. Deflecting is not the warmer experience. It simply was not the count that tracked the outcome.
The reasons to hold this loosely
None of the ten studies could assign anyone a real secret, and the authors say so: the data cannot demonstrate causality, and they doubt an experiment could ethically manufacture a secret from a partner. Every number here is two things a person estimated about themselves, and the authors are open that recall of something as fugitive as a passing thought is imperfect. The well-being measure in the early studies was a single question. In the significant-secret study, the life-satisfaction result cleared the usual threshold by a hair.
Two later publications complicate it further, and Slepian’s name is on both.
The first is about direction. Writing in 2024, Slepian notes that being preoccupied with a secret you are keeping from a partner goes with a worse relationship, and his own reading of work by other researchers is that the preoccupation “is a symptom of low-quality relationships rather than a cause.” The arrow may run the other way entirely: unhappy relationships may generate the kind of secret that will not leave you alone. He is blunt that establishing causation here “will be highly difficult in this domain.”
The second is about the numbers themselves. In a paper published online in 2024 and in print in 2025, a team including Slepian counted close to the time the moments happened rather than asking people to recall a month. One study had people fill in a diary each evening for a week; a second pinged them six times a day. Together they captured more than 4,600 moments. The ordering survived: people still reported thinking about their secrets more than concealing them. The magnitudes did not. Measured this way, rates of thinking about a secret came out four times higher than previously published estimates in the first study and almost eight times higher in the second, or three and six times once outliers were handled. The authors add that this does not establish which method is right, only that the newer numbers and the older ones do not match.
So the honest version is smaller than the headline. The ordering looks robust across methods; the specific tallies do not; and whether the thinking does anything to you, or merely reports on something already wrong, is unsettled. We read studies for a living and are not clinicians, and none of this is a diagnosis or a substitute for one. A secret that is sitting on your chest, particularly one involving harm or your own safety, belongs with a therapist rather than with a literature review; where the secret is from a partner, that is couples-therapy territory. The 2017 paper tests no fix. Its authors go only as far as suggesting that approaches already known to reduce mind-wandering, such as acceptance- or mindfulness-based therapies, might help people who cannot tell anyone, and they offer that as a direction, not a result.
What it is good for instead
What it can do is relocate the problem, which is not nothing.
If you have been bracing for the conversation, you have been bracing for the rarer event. The research consistently suggests the larger share of the experience is unscheduled, and that it happens more often out of the other person’s sight than in it. What is known about relief points the same way: confiding in a third party, rather than confessing to the person the secret is kept from, is the better-studied of the two and generally goes better than people expect. One exception the 2024 review records is rare and specific: a confidant who responds very badly, which predicts lower well-being for the person who spoke.
It also suggests that a secret is not a performance with an audience. It is closer to an unfinished errand, and unfinished errands are exactly the sort of thing the mind is thought to keep near the surface, which is one of four explanations the authors offer, none excluding the others. The plainest of them is that a day contains far more minutes alone with your own head than minutes across a table from the person you are keeping it from.
So the kitchen is quiet and it arrives again. That moment is not an interruption of the secret. That moment is where the secret mostly lives.
