The psychology of shared meals at work

Eating together tends to make people more cooperative and more trusting, and the best experimental evidence points to something specific: sharing dishes, and eating the same food, appears to matter more than simply being in the same room with lunch. Most of that evidence comes from lab studies, often with strangers. The claim you'll often read, that teams who eat together perform better at work, rests on much thinner ground.

Here's what the research shows, what it doesn't, and how to use it when you plan a team lunch or dinner.

Sharing a plate makes people coordinate

The strongest work comes from Kaitlin Woolley and Ayelet Fishbach. In a 2019 paper in Psychological Science, they ran a series of experiments with 1,476 participants. People who ate from a shared plate, in the style of a Chinese family meal, then cooperated more in negotiation games and social dilemmas than people who ate the same food from separate plates. The effect held among strangers, not only friends.

Their explanation centers on coordination. Taking from a shared dish means watching what others take, waiting your turn and making sure everyone gets enough. In their studies, that habit of paying attention to the other person carried into the task that followed.

A 2020 follow-up in the same journal by Cao, Kong and Galinsky tested a more realistic setup: 97 pairs of students negotiating a job offer with several issues on the table, talking freely. Pairs who snacked from a common bowl reached agreements that created more value for both sides than pairs who ate from separate bowls or didn't eat. Those who ate separately did no better than those who didn't eat at all.

Eating the same food builds trust

An earlier study by the same team, published in the Journal of Consumer Psychology in 2017, found that strangers assigned to eat similar foods trusted each other more in a trust game, and cooperated more in a simulated labor negotiation, than strangers eating different foods. Food choice seems to work as a quick signal of "we're alike."

For a team lunch, that suggests a few shared dishes do more than a dozen separate orders, with an important exception covered below for allergies and dietary needs.

Eating together is linked to feeling connected

Outside the lab, the best-known study is Robin Dunbar's 2017 paper in Adaptive Human Behavior and Physiology. It used a survey of 2,000 UK adults, run by the polling firm OnePoll in April 2016 in collaboration with the Big Lunch community-meal project. People who ate with others more often reported being happier and more satisfied with life, more trusting, and having more friends they could rely on. Meals that left people feeling closer to their companions involved more people, more laughter and reminiscing, and alcohol.

Read this one carefully. It's a one-time survey, so it shows a link, not a cause; Dunbar's path analysis suggests eating together leads to closeness rather than the reverse, but a single survey can't settle that. It's also about meals with friends and family, not coworkers.

What we left out, and why

You'll find two claims in many articles on this topic that we don't repeat.

The first is the Cornell firefighter study (Kniffin, Wansink and Sobal, 2015), which linked eating together at the firehouse with better team performance. The paper itself hasn't been retracted, but co-author Brian Wansink left Cornell in 2018 after a university investigation found research misconduct and many of his papers were retracted. Until someone replicates it independently, we don't lean on it.

The second is the idea that a shared meal releases oxytocin and "wires" coworkers to like each other. We found no good human evidence for that in workplace meals.

We also couldn't trace two figures from an earlier version of this post: a "36% higher collaborative performance" for teams that eat together three times a week, and "62% of knowledge workers eat lunch alone." Both are gone.

Planning a team meal around the evidence

The table below pairs each choice with what a study found, where one exists. The recommendations are our guidance.

Choice	What a study found	Our guidance
Family-style or individual plates	Shared plates increased cooperation in experiments (2019, 2020)	Order family-style when the menu allows; a few shared starters work if mains are individual
One menu or many	Eating similar food increased trust between strangers (2017)	A set menu or a few shared dishes, plus labeled options for dietary needs
Seating	Not tested for meals	Tables of four to six, or a long table with people across from each other, so everyone can join one conversation
Length	Meals that felt closer involved laughter and reminiscing (2017)	60 to 90 minutes, so conversation gets past small talk
Timing	No direct evidence	During work hours, midweek, so people don't trade personal time to attend
Alcohol	Linked to feeling closer in Dunbar's survey	Optional, never central; plenty of good non-alcoholic choices
Attendance	No direct evidence	Voluntary. A required meal is a meeting with food
Frequency	No study compares weekly with monthly team meals	Start monthly; go weekly if people keep coming

When shared plates are the wrong call

Family-style service carries real risk for anyone with a food allergy or celiac disease, because serving spoons move between dishes. If someone on the team has a serious allergy, give them their own plate served first, label every dish, and keep a separate serving utensil for each one. Our guide on handling dietary restrictions at team events covers how to ask and what to tell the restaurant.

Making it a habit

One lunch is pleasant. A regular one gives people repeated time together, which is how colleagues turn into people they trust. If you're starting from nothing, pick a recurring slot, a place within walking distance and a budget you can sustain. How to plan a team dinner has formats and per-person ranges, and the data behind team bonding puts meals alongside the other kinds of team events with evidence behind them.

Where TeamOutings fits

In TeamOutings, a team lunch is an outing: set the date, pick the restaurant from saved locations, note the cost per person, and collect Yes, No and Maybe replies with a live headcount you can give the restaurant. A poll can settle where to go, the outing's chat handles "is there parking?", and the photo album keeps the pictures in one place.

TeamOutings doesn't collect dietary needs or send reminders, so ask about allergies directly and follow up with people yourself.