A shared meal is such an ordinary event that it barely looks like data. Someone passes the salt. A story gets interrupted. Two people sit at the same table without necessarily saying anything profound.
Yet when researchers examined responses from more than 150,000 people, the number of lunches and dinners they ate with others tracked their wellbeing remarkably closely. In statistical terms, meal sharing predicted life satisfaction about as well as income or employment status did.
The other half of the finding is harder to brush aside. In 2023, 26 per cent of Americans said they had eaten every meal alone on the previous day. That was 53 per cent higher than the corresponding figure in 2003.
Those numbers deserve attention, but also a careful reading. This is one research programme built from two observational datasets, not proof that sitting at a table causes happiness. The global survey and the American time-use study also asked different questions. Their results fit together, but they are not interchangeable.
What the researchers actually asked
The main analysis was conducted by Jan-Emmanuel De Neve of the University of Oxford with Andrew Dugan of Gallup, Micah Kaats of Harvard Kennedy School and Alberto Prati of University College London. So while Oxford led the work, describing it simply as an Oxford survey compresses a four-institution collaboration.
For the global part, Gallup asked people in 142 countries and territories how many days in the previous week they had eaten lunch with someone they knew, and how many days they had eaten dinner with someone they knew. The questions appeared in the 2022 and 2023 Gallup World Polls.
In the researchers’ peer-reviewed analysis in Scientific Reports, people who shared more meals generally reported higher life evaluations and more positive emotions, as well as fewer negative emotions. The association remained after the team accounted for age, gender, education, employment, income, household size, food affordability and differences between countries.
That makes meal sharing a useful social indicator. It does not turn it into a treatment. Happier people may be more inclined to arrange meals, strong relationships may produce both shared meals and wellbeing, or the meals themselves may help. The data cannot tell us how much each explanation contributes.
“As strong as income” describes prediction, not a causal effect
The most eye-catching comparison in the study is also the easiest to misunderstand. The researchers ran separate statistical models and asked how much variation in people’s life evaluations could be explained by meal sharing, income and unemployment. On that measure, the number of shared meals performed about as well as the established economic indicators.
It does not mean that one dinner has the same effect as a pay rise. It does not establish that arranging another lunch will lift someone’s life satisfaction by a fixed amount. It means the simple meal question contains an unexpectedly large amount of information about how people say their lives are going.
The steepest difference appeared at the bottom of the scale. People reporting no shared lunches or dinners in the previous week rated their lives at an average of 4.9 out of 10. Those reporting just one shared meal averaged 5.2. Across countries, each additional shared meal per week was associated with roughly 0.2 points in life evaluation, although the pattern was not identical everywhere.
I find that comparison more useful when it is kept modest. As I noted in an earlier piece about money and happiness, income matters, but it never operates in a social vacuum. A meal-sharing question may be picking up some of that surrounding texture: time, companionship, routine and the feeling that there is someone with whom an ordinary hour can be spent.
The American “one in four” figure comes from a different dataset
The US trend was not calculated from the global Gallup questions. It came from the US Bureau of Labor Statistics’ American Time Use Survey, which asks people to reconstruct what they did during the previous day.
The researchers analysed records from 234,185 American adults collected between 2003 and 2023. They classified a person as dining alone if every eating occasion in that day’s diary took place without anyone else present. In 2023, 26 per cent fell into that group, up from about 17 per cent in 2003.
This is where wording matters. The result does not show that one quarter of Americans habitually eats every meal alone, every day. It shows that one quarter did so on the particular day captured by the survey. Because the survey is designed to represent the population across many diary days, it can reveal a national trend, but it cannot turn a one-day record into a permanent description of an individual’s life.
Living alone explains only part of the rise
It would be tempting to attribute the change entirely to more single-person households. The analysis suggests that is only a small part of the story. Changes in living arrangements accounted for about 17 per cent of the overall rise in eating alone.
Among Americans who lived with other people, the proportion eating every meal alone on the diary day climbed from 12 per cent to 18 per cent. Among those living alone, it rose from 55 per cent to 69 per cent. Solitary eating therefore increased inside shared homes as well as outside them.
Younger adults recorded some of the largest relative increases. That does not reveal why. Work schedules, commuting, phones, study, food delivery and changing household routines are all plausible parts of the background, but the time diaries do not isolate a single cause.
The distinction matters because “eating alone” is not automatically the same thing as loneliness. A quiet breakfast can be restorative. A crowded family dinner can feel tense. The researchers were measuring the presence of other people, not the quality of the encounter.
The Nordic pattern is not one tidy happiness lesson
The country rankings resist a neat cultural slogan. Latin America and the Caribbean led the world in meal sharing, while South Asia recorded the lowest regional average. Within northern Europe, Iceland ranked ninth globally with about 10 shared lunches and dinners per week. Sweden ranked 19th, Denmark 33rd and Norway 37th in the report’s country appendix.
Those are relatively high figures, but they do not support the claim that Nordic happiness is simply the product of communal dining. Iceland is Nordic rather than Scandinavian, the countries differ from one another, and national averages conceal a great deal of variation between households and individuals.
The broader World Happiness Report chapter is more interesting than that. It treats meal sharing as a comparable measure of social connection across cultures, while acknowledging that a shared meal may carry different meanings in Bogotá, Stockholm, Mumbai or a small American town.
A shared meal is a marker, not a prescription
The new analysis is not the first to notice that eating together and social connection travel together. A separate 2017 study of social eating, led by Oxford psychologist Robin Dunbar, also found that people who ate socially more often reported stronger bonds and greater engagement with their communities. That earlier work was observational too, so it adds context rather than causal proof.
There are also practical differences hidden inside the phrase “share a meal”. For one person it means a relaxed dinner with friends. For another it means care work, expense or an obligation performed after a long shift. Some people lack company; others deliberately protect time alone. A responsible reading leaves room for all of that.
What the researchers have identified is less a rule for living than a surprisingly revealing question. Income and employment tell us something important about the resources around a person. Shared meals may tell us something about whether everyday life regularly overlaps with somebody else’s.
A meal records a simple fact: whether another person was there. That modest detail is what makes the measure useful.