How couples meet: what the statistics really measure
39% of new US couples met online, or 10%, or 17% in Switzerland: each figure is right about a different question. Here is what each study measured.
The best known statistic on how couples meet comes from Stanford: among American heterosexual couples who met in 2017, 39% met online, more than through friends (20%), and online has been the most common route since about 2013. It is a sound figure, but it answers a narrow question: new couples, in one country, in one year, as described in their own words. Pew Research Center, counting all partnered American adults in 2022 however long ago they met, found that one in ten met their current partner through a dating site or app. A Swiss government survey from 2018, as analyzed by Gina Potarca, puts online meeting at about 17% of people who met their partner in the previous ten years. All three are correct. They measure different things, and the useful skill is knowing which question a number answers before you repeat it.
Three studies side by side
| Study | Who was asked | When | How the question worked | Headline |
|---|---|---|---|---|
| HCMST 2017, Stanford | nationally representative sample of English-literate US adults; analysis of heterosexual couples | 2009 and 2017 surveys, pooled; couples by year they met | open text, the story of how you met, coded by researchers; more than one code allowed | 39% of couples who met in 2017 met online |
| Pew Research Center | 6,034 US adults | July 5 to 17, 2022 | partnered adults asked how they met their current partner; we read Pew's summary, not the questionnaire | one in ten met their current partner on a dating site or app |
| Family and Generations Survey, Swiss Federal Statistical Office | sample of Swiss residents aged 15 to 79; 3,245 who met their partner in the last ten years | 2018 | one answer from nine options | 4.12% apps, 7.33% dating websites, 5.24% other online |
What Stanford actually measured
The How Couples Meet and Stay Together survey, run by Michael Rosenfeld, Reuben Thomas and Sonia Hausen, asked respondents to "write the story of how you and [partner name] first met and got to know one another", and to say how and where. The researchers then coded each story into categories. A couple who met online and had their first date in a bar counts under both, which is why the percentages do not add up to 100.
Their 2019 paper in PNAS compares smoothed values for heterosexual couples who met in 1995 and in 2017:
| How they met | 1995 | 2017 |
|---|---|---|
| Online | 2% | 39% |
| Through friends | 33% | 20% |
| In a bar or restaurant | 19% | 27% |
| Through or as coworkers | 19% | 11% |
| Through family | 15% | 7% |
| In grade school or high school | 10% | 5% |
| In college | 9% | 4% |
| Through or as neighbors | 8% | 3% |
| In church | 7% | 4% |
The bar and restaurant row looks like a revival of nightlife. It is not. The authors explain that the post-2010 rise is due entirely to couples who met online and then had their first meeting in a bar or restaurant. Exclude them and meeting in bars had been declining since 1995.
The paper's second finding is about friends. In the 2009 survey, 11.2% of couples who met online had been put in touch by a friend or another third person; in 2017 it was 3.7%. Online meeting is not only growing, it is replacing the introductions that friends and family used to make, which is what the title, disintermediating your friends, refers to.
Two limits are worth stating. The 2017 survey had a response rate of 60%, and once earlier stages of panel recruitment are included the cumulative rate is 11%, which the authors report themselves. And the 39% is a share of couples who formed in 2017, not of everyone in a relationship. Stanford's data page adds that the 2017 sample of 3,510 people was followed up in 2020 and 2022, so the series continues.
Why Pew's figure is so much lower
Pew's 2022 survey asked partnered adults, meaning married, living with a partner or in a committed relationship, whether they met their current partner through a dating site or app. One in ten said yes. Among partnered adults under 30 it was one in five.
Much of the gap with Stanford is arithmetic. Pew's question covers every existing couple, including people who met in the 1970s, before dating sites existed. Stanford's 39% covers only couples formed in a single recent year. A stock of all couples changes slowly; the flow of new couples changes fast. Pew also found that 30% of American adults have ever used a dating site or app, rising to 53% of those under 30, which is a measure of use, not of couples formed.
The European picture
Nothing like HCMST exists for Europe as a whole, or for the UK, as far as we could find. The closest careful study is Gina Potarca's analysis of the 2018 Family and Generations Survey, run by the Swiss Federal Statistical Office as part of the national census program, with telephone interviews followed by a questionnaire. Of 36,029 people sampled, 16,815 took part. The 2018 round was the first to ask where couples met.
The question was "How did you meet your partner?", with one answer allowed from nine options, among them friends, family, school or work, a bar or party, a hobby or club, a dating website, a dating app, and a social network or other online service. Among the 3,245 people who had met their partner in the previous ten years, 4.12% named an app, 7.33% a dating website and 5.24% another online service, all weighted. Our addition: about 17% online in total.
That figure is not comparable with Stanford's 39%. It covers ten years of couples rather than one, it allows a single answer where Stanford allowed several, and Switzerland is not the United States. Potarca also found no difference in relationship or life satisfaction between couples formed through apps and couples formed elsewhere, and stronger intentions to live together among the app couples.
Do dating apps actually work? What the numbers say
It depends what "work" means, and the studies above answer three versions of the question.
Do couples form there? Yes, in large numbers. Rosenfeld and colleagues date the point at which meeting online overtook meeting through friends, for American heterosexual couples, to around 2013, and Pew found that one in ten partnered US adults, and one in five under 30, met their current partner on a dating site or app.
Do those couples last? As well as other couples, on the best evidence we found. The PNAS paper reports that follow-ups of the 2009 HCMST sample showed neither breakup rates nor relationship quality depended on how couples met, and that in the 2017 data how couples met had no significant effect on the risk of breaking up once the decade of meeting was controlled for. Potarca's Swiss data point the same way on satisfaction.
Do users feel it worked? Only just more than half. In Pew's 2022 survey, 53% of Americans who had ever used a dating site or app called their experience positive and 46% negative. Men were more positive (57%) than women, who split 48% to 51%.
None of these studies says which app works best or how long anyone searched first. A route that produces many couples can still leave most of its users without one.
Dating app statistics: who uses them in the US and UK
In the US, Pew's 2022 survey found men somewhat more likely than women to have ever used a dating site or app (34% against 27%), and never-married adults far more likely than married ones (52% against 16%). It found no statistically significant difference by race or ethnicity: similar shares of White, Black, Hispanic and Asian adults had tried one. Among online daters, 46% had used Tinder, 31% Match and 28% Bumble.
The UK has no survey of the same design, but it has measurement. Ofcom's Online Nation 2025 report, based on Ipsos iris audience data, found that 11% of UK online adults visited an online dating service in May 2025, about the same share as in each of the three previous years. Visitors were 68% men, 3.6 million of them. Tinder and Hinge each reached 3.1% of online adults, about 1.5 million people, and Bumble 2.3%; Tinder's female audience fell from 735,000 to 384,000 in a year. These are visits in one month, not answers about ever using an app, so the 11% cannot be set against Pew's 30%.
How to read the next statistic you see
- Which population? One country, all adults, or only people in a relationship?
- Stock or flow? Everyone who is now a couple, or couples who formed in a particular year?
- What was asked? An open story, a list with one answer, or a list allowing several?
- When? A figure from before the pandemic, or before apps became common, may describe a different market.
A number that cannot answer all four is not wrong, but it is not yet a fact about how couples meet.
Frequently asked questions
What share of couples now meet online?
Is that figure true outside the United States?
How do researchers measure how couples met?
Are couples who met on apps different?
Does online dating actually work?
Sources
- Rosenfeld, Thomas and Hausen: Disintermediating your friends, PNAS 116 (2019) Read October 7, 2026, in full through PubMed Central (PMC6731751). Supports the abstract's finding that meeting online overtook meeting through friends around 2013; Table 1 (heterosexual couples, 1995 against 2017: online 2% to 39%, through friends 33% to 20%, family 15% to 7%, coworkers 19% to 11%, bar or restaurant 19% to 27%, school 10% to 5%, church 7% to 4%, neighbors 8% to 3%, college 9% to 4%; N = 5,421; smoothed values); the open-ended question wording; categories not mutually exclusive; the explanation of the bar and restaurant rise; third-person mediation among online couples falling from 11.2% to 3.7%; response rates of 60% in 2017 and a cumulative 11%. Rechecked October 16, 2026 in the PMC full text: the statement that HCMST 2009 follow-ups showed neither breakup rates nor relationship quality were influenced by how couples met, and that the HCMST 2017 retrospective breakup data showed no significant effect of how couples met on the hazard of breakup once decade of meeting was controlled for.
- Stanford University Libraries: How Couples Meet and Stay Together 2017 (HCMST 2017) Read October 7, 2026. Supports the 3,510 respondents with no overlap with the 2009 survey, the follow-ups in 2020 (2,107) and 2022 (1,722), the new questions on phone apps, and the coding of open-text answers by the investigators.
- Pew Research Center: Key findings about online dating in the U.S. (February 2, 2023) Read October 7, 2026, and again on October 8, 2026. Supports the survey of 6,034 adults from July 5 to 17, 2022, the 30% who have ever used a dating site or app (53% of under 30s), and the finding that one in ten partnered adults met their current partner on a dating site or app, one in five among partnered adults under 30. Reread October 16, 2026. Also supports ever use by sex (men 34%, women 27%) and marital status (never married 52%, married 16%), no statistically significant difference by race or ethnicity, Tinder (46%), Match (31%) and Bumble (28%) as the most used among online daters, and 53% of users calling their experience positive against 46% negative (men 57% positive, women 48% positive and 51% negative).
- Potarca: Couples who met through dating apps in Switzerland, PLOS ONE (December 30, 2020) Read October 7, 2026, in full on journals.plos.org. Supports the data source (Swiss Federal Statistical Office, Family and Generations Survey 2018, telephone interviews, 16,815 of 36,029 sampled people taking part), the single-answer question and its nine options, the analysis sample of 3,245 people who met their partner in the previous ten years, the weighted shares meeting through apps (4.12%), dating websites (7.33%) and other online services (5.24%), and the findings on satisfaction and cohabiting intentions.
- Ofcom: Online Nation 2025 report (December 10, 2025) Read October 16, 2026, dating section and Figure 39. Ipsos iris online audience measurement, UK online adults aged 18 and over. Supports 11% of UK online adults visiting an online dating service in May 2025 (10% to 11% in the three previous Mays), visitors 68% male (3.6 million), Tinder and Hinge each at 3.1% reach (1.5 million), Bumble 2.3% (1.1 million), and Tinder's female audience falling from 735,000 in May 2024 to 384,000 in May 2025.