Situationship Statistics: What Has Actually Been Measured

The number everyone quotes, and the survey it came from

Almost every article on this subject opens with a share — a third of people, half of young adults, some figure presented as settled. There is a real number underneath most of them, and it is worth knowing exactly what it is and what it was asked.

In January 2024 YouGov surveyed 1,110 US adults over two days and gave them a definition before asking anything: a situationship is a romantic connection in a grey area, neither strictly platonic nor officially a committed relationship. 39% of respondents said they had been in one. Among 18- to 34-year-olds, 50%.

That is the source of the number. It is a real, weighted, nationally representative survey, and the figure is not in dispute. What it can support is narrower than how it usually gets used.

It is a lifetime prevalence question — have you ever — not a snapshot of how many people are in one now. Those two quantities are very different, and the second is the one most readers think they are being told. A person who spent four months in an undefined arrangement in 2016 counts in that 39% forever.

It also depends on people accepting a label. The survey supplied the definition, which helps, but respondents still had to recognise their own history in it. There are reasons to think that recognition is uneven: the term is much more current among younger respondents, and a 58-year-old describing a two-year period of seeing someone without ever discussing what it was may simply not file that under a word they first heard last year. Whether that pushes the true figure up or down is not knowable from one survey.

The rest of what YouGov published is about how these connections start — 55% in person, 17% through social media (29% among 18- to 34-year-olds), 7% through dating apps. That last share is lower than most people expect, and it sits oddly beside the fact that online is now the most common route into a relationship that does get defined. Note what is not there: nothing about how they ended, how long they lasted, or how the people in them felt. The survey counted them. It did not follow them.

That pattern — a good prevalence number and nothing after it — repeats across everything measured on this subject, and it is most of what this page is about. The same shape shows up wherever a dating claim gets repeated without a study behind it, which is the subject of the rules with numbers in them that nobody can source.

The survey that asked the better question

A second survey is more useful, and it is almost never the one quoted, probably because it never uses the word.

In Romantic Recession, published in January 2025, Daniel A. Cox and Kelsey Eyre Hammond of the Survey Center on American Life reported results from the American Gender and Society Survey — 5,837 US adults, fielded 16-26 August 2024. Instead of naming a category and asking whether people had been in one, it asked about the experience directly: had they had a sexual or romantic relationship with someone and been unsure how they would define it.

36% of Gen Z adults said yes — 38% of Gen Z women, 34% of Gen Z men. Among baby boomers, 23%, with men and women at roughly the same rate. Millennials and Gen Xers came in close behind Gen Z rather than between the two extremes; the report does not give their exact figures.

Two things in that are worth slowing down for.

The first is the size of the sample. 5,837 respondents is more than five times the YouGov survey, which matters for the generational breakdowns: cutting 1,110 people by both age and gender leaves cells small enough that a few percentage points of movement means very little.

The second is the wording. Asking whether someone has been unsure how to define a relationship describes an experience. Asking whether someone has been in a situationship asks them to sort their history into a category. The first travels across generations; the second does not, which is exactly why the boomer figure of 23% is more interesting than it looks. Nearly a quarter of a generation that had no word for this reports having lived it.

Which is the finding that most of the genre gets backwards. The vocabulary is new. The experience is not, and the gap between 36% and 23% is real but a good deal smaller than the story about a uniquely modern condition needs it to be. The wider version of that argument — what has actually changed in dating and what only sounds like it has — is in the real numbers behind modern dating.

What Pew has actually asked, and what this page got wrong

A correction, and it belongs here rather than at the bottom. Until 10 September 2026 this page opened by reporting that Pew Research Center had surveyed Americans in 2023 and found that roughly 1 in 4 adults under 30 had been in a relationship neither person had defined, that about 38% of recent daters described a relationship as ambiguous, and that people gave specific reasons for not defining things at rates of 41%, 37% and 29%. It also reported that Pew had found that about 1 in 5 of the people who preferred ambiguity had been hurt by a previous undefined relationship.

Pew published none of those figures. They were not misread from a Pew report, they were not taken from the wrong year, and they do not correspond to a question Pew has asked. The correct statement is the plain one: Pew Research Center has not measured situationships. Every number in that section has been removed, and the change is logged on our corrections page.

What Pew has measured, and it is genuinely useful, is what single people say they want. In From Looking for Love to Swiping the Field (published February 2023, fieldwork 5-17 July 2022), 42% of single US adults said they were looking for a committed relationship and/or casual dates, down from 49% in 2019. That drop is the headline finding and it points somewhere real: a majority of single Americans were not, at that moment, looking for either.

The age gradient is steep. 57% of unpartnered adults under 50 said they were looking, against 36% of those aged 50 to 64 and 16% of those 65 and older. Among people using dating sites and apps, 44% said meeting a long-term partner was a major reason for it and 40% said casual dating was.

None of that is a situationship statistic, and the distinction is the whole point of this section. Pew asked what people are looking for. YouGov and the Survey Center asked what people have been in. Those are different questions, and the reason the fabricated version of this page was so easy to write is that a Pew citation makes any number sound settled. That is what our editorial policy exists to prevent, and it did not prevent this one.

The research literature does not study situationships. It studies relational uncertainty.

Surveys count. They do not explain. For explanation you have to leave the polling and go into a communication-research literature that has been working on this since the late 1990s and has never once used the word.

The construct is relational uncertainty, and the instrument that made it measurable is Knobloch and Solomon's 1999 paper in Communication Studies. Their contribution was to stop treating uncertainty as one fog and split it into three sources:

Self uncertainty — how sure you are about your own involvement. Do I want this. How much do I want it.

Partner uncertainty — how sure you are about theirs. The one most people mean when they say they do not know where they stand.

Relationship uncertainty — how sure you are about the relationship as a thing in itself, which they broke down further into uncertainty about behavioural norms (what is allowed here), mutuality (whether the feelings run both ways), definition (what this is), and future (where it goes).

That taxonomy is doing more work than it looks like. The popular framing of an undefined relationship collapses all of this into a single question — what are we — when the measurement research says these vary independently. Someone can be entirely certain of their own feelings and their partner's and still have no idea what the arrangement is called or what it permits. Someone else can have an agreed label and be deeply unsure of the feelings underneath it. Those are different situations with different remedies, and a survey that asks whether you have ever been in a situationship cannot tell them apart.

It also explains why the advice genre keeps prescribing the same conversation for everything. If the uncertainty is about definition, naming it resolves it. If it is self uncertainty, naming it resolves nothing, because the missing information is not in the other person. That distinction is the useful part of this literature and it almost never survives the trip into an article.

The related question of what a relationship's own ingredients predict — satisfaction, alternatives, investment, commitment, and how they come apart — has a separate and much larger literature behind it, which is what our relationship satisfaction test is built on.

What happened when they tested the obvious prediction

Here is where this gets more honest than the genre usually allows, and it is the reason this literature is worth reading rather than summarising.

Solomon and Knobloch had a clean theoretical prediction: the transition from casual to serious involvement is the turbulent stretch of a courtship, so relational uncertainty should peak at moderate intimacy — high in the middle, lower at both ends. It is intuitive. It matches how people describe the experience. It is the shape any writer would assume.

They tested it in 2001 on 341 people in dating relationships, measuring intimacy, relationship uncertainty, and how much a partner interfered in everyday activities.

The prediction failed. Rather than a peak in the middle, they found a negative linear association: uncertainty simply declined as intimacy rose, with no turbulent middle at all. The curvilinear pattern they predicted did show up for interference from partners — greatest at moderate intimacy, as expected — but the authors report the effect size as small.

Three years later they tested the model again, this time on how people appraise irritating things a partner does. Negative appraisals were curvilinearly related to intimacy, and the effect size was again small. Relational uncertainty and interference were both positively associated with negative appraisals, as predicted. But — and they say so plainly — neither one mediated the curvilinear relationship, which was the mechanism the whole model proposed.

Two studies, two partial failures, both reported by the researchers who made the predictions. That is what a functioning literature looks like, and it is the opposite of how this topic is written about elsewhere, where the mechanism is always confirmed and always confidently.

What survives is more modest than the theory wanted: uncertainty and negative appraisal travel together, the effects are small, and the tidy story about a turbulent middle phase of courtship did not hold up when it was measured. Anyone telling you that the ambiguous stage is scientifically the hardest stage is telling you something the two studies designed to show it did not find. The same pattern — a confident claim, a study underneath it that says something narrower — runs through what questions can and cannot reveal about compatibility.

What uncertainty does track

Set the failed predictions aside and there is a stable body of findings about what relational uncertainty co-occurs with.

In 2003 Knobloch and Solomon asked 141 people in dating relationships to describe events that had made them more certain or less certain about the relationship, and what they felt and did afterwards. Events that increased certainty coincided with relatively high happiness. Events that increased uncertainty corresponded with heightened anger and sadness. Integrative responses — talking it through rather than withdrawing or attacking — were a common reaction to both kinds of event, which is a more encouraging finding than the framing usually allows.

The most complete picture comes from a 2020 meta-analytic review in Human Communication Research. Goodboy, Bolkan, Sharabi, Myers and Baker ran 27 separate meta-analyses across nearly two decades of relational turbulence research. Twelve of them (k = 9-15, n = 1,395-5,493) confirmed that the model's variables — self uncertainty, partner uncertainty, relationship uncertainty and partner interference — correlate on average with topic avoidance, depressive symptoms and relationship satisfaction. A further fifteen (k = 4-41, n = 930-8,975) pooled the correlations among the variables themselves, and a structural model fitted to that pooled matrix fit well.

So the association with distress is real and it has been pooled across a large evidence base. Two cautions on reading it, both of which the meta-analysis itself does not remove.

The first is effect size. The abstract reports that the correlations held on average; it does not license the sentence people want to write, which is that ambiguity makes you depressed. Correlated is not the same as large, and the primary studies above were reporting small effects.

The second is topic avoidance, which sits in that list of outcomes and is the most practically interesting item in it. The variable that travels with uncertainty is not only how bad you feel — it is how much you stop raising. That is a loop with an obvious shape: uncertainty makes the subject harder to bring up, and not bringing it up keeps the uncertainty in place. Nobody has established which end starts it.

The direction of the arrow, which nobody has established

Every study in the previous two sections is cross-sectional. People were measured once, and the associations are between things observed at the same moment.

That leaves the question everyone actually has completely open. Does ambiguity make people miserable, or do people who are already struggling end up in ambiguous arrangements more often? Both stories fit the data equally well, and a third fits it too: some prior thing — an anxious period, a recent ending, a partner who avoids commitment generally — could produce both the uncertainty and the low mood without either causing the other.

There is no study that resolves this, and it is worth being blunt that the absence is not a technicality. It is the difference between "define the relationship and you will feel better" and "people who feel better are more likely to define the relationship," and every practical recommendation on this subject depends on which one is true.

Three more things nobody has measured, all of which get asserted constantly:

How situationships end. Neither survey followed anyone. There is no published figure for what share become committed relationships, what share end, or what share simply dissolve without either.

How long they last. Widely quoted durations circulate with no source under them.

Whether the people in them wanted them. The Survey Center asked whether people had been unsure how to define a relationship. It did not ask whether they minded. The report gives prevalence by generation and gender, and stops.

Which means the confident version of this topic — that situationships are a modern trap, that one person always wants more, that they rarely convert — is a set of claims about outcomes built entirely on surveys that measured prevalence. The claims may well be right. They have not been tested. If you have been in one and are trying to work out whether it is actually over, the signs that you have moved on, or have not is about that question rather than this one.

Where the undefined stops being undefined

This distinction has nothing to do with the research above and it goes in anyway, because the two situations look similar from outside and are not remotely the same thing.

An undefined relationship, in the sense everything on this page has been about, means the status is unresolved for both people. Neither has a settled answer. That is ordinary, it is common across every generation measured, and it resolves in whatever direction it resolves.

Something else is going on when the ambiguity is maintained on you. Named plainly, because vagueness here helps nobody:

Enforced secrecy — you are told not to mention the relationship, not to post about it, not to tell mutual friends, and there are consequences when you do.

Isolation as a condition — meeting their friends or family is permanently deferred, and the undefined status is the stated reason.

Monitoring — your phone, your location or your movements are checked, or you are expected to account for your time, while the relationship is officially nothing.

Escalation at a direct question — asking what this is reliably produces anger, withdrawal as a penalty, or a threat to end contact.

The fourth is the one people most often explain away, and it is the clearest. In an ordinary undefined relationship a direct question is uncomfortable. It does not carry a cost imposed by the other person.

In the United Kingdom, controlling or coercive behaviour in an intimate or family relationship is a criminal offence, and the Home Office statutory guidance lists behaviours of exactly this kind. That guidance is in the sources at the end. So is Find A Helpline, which lists free confidential services by country; in the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day.

Nothing on this page can assess a particular relationship, and no article can. If what is described in this section is familiar, the useful next step is a person, not a page. Related patterns and where the line falls are covered in the psychology of jealousy.

What the numbers are actually good for

After all of that, the honest answer to what the statistics can do for anyone in this situation is: less than the headline promises, and not nothing.

What they cannot do is tell you what to do. Knowing that 39% of US adults have been through something like this, or that 36% of Gen Z have been unsure how to define a relationship, gives you no information about your own, because prevalence says nothing about any individual case. A common thing can still be a bad idea for you; an uncommon one can be exactly right.

What they can do is remove a false premise. The consistent thing people report about undefined relationships is a sense that theirs is a personal failure — a misjudgement, a weakness, an inability to ask a simple question. The surveys are a straightforward answer to that: this is somewhere between a quarter and half of adults depending on how you ask, it spans every generation measured, and 23% of baby boomers lived it without a word for it. Whatever it is, it is not evidence of something wrong with you specifically.

And there is one finding worth carrying out of the research literature, stated at the size it actually is. Topic avoidance travels with relational uncertainty. Not causally, not established in any direction, but reliably enough to survive pooling across a large evidence base. The thing that tracks with not knowing where you stand is not asking. That is not advice, and it is certainly not a script for producing a particular answer from another person. It is a description of a loop that a lot of people are in, and noticing you are in it is worth more than another statistic.

Whether the pattern repeats for you across different people is a separate and more useful question than any of this — why intelligent people keep choosing badly and the attachment style test are both about that. The practical companion to this page, on recognising the arrangement from the inside rather than from survey data, is the signs, and what each one is evidence of. More on this subject under Dating.