From your workplace to your home, life’s big, elusive secret often seems to be how to get people working together without killing each other. Hundreds of think-pieces exist on how to be a good manager; even more exist on how to be a good parent. (Some speak to both.) But few explore the actual conditions necessary for building effective teamwork.

A new study out of The Ohio State University sheds some light on that. The key to getting people to work together effectively, the researchers conclude, could be giving them the flexibility to choose their collaborators and the comfort of working with established contacts.

But first, it’s about managing expectations. Cooperation between humans makes no sense, explained David Melamed, lead author of the study, published in the journal Proceedings of the National Academy of Sciences.

“From an evolutionary perspective, cooperation shouldn’t exist between people — you always do better by not cooperating because then people can’t rip you off or take advantage of you,” Melamed said. “Especially in a one-time interaction, it’s essentially paying a cost for someone else to benefit, and researchers have been working for a long time to understand why people evolved to work together.”

In this study, Melamed and his co-authors aimed to uncover what conditions led people to collaborate most willingly. They recruited 810 people to play online games in which each player started out with 1,000 monetary units that translated to $1 in real money that they could pocket. If one player agreed to pay another player 50 monetary units, that second person would actually acquire 100 units.

“So, if you essentially agreed to give up five cents, someone else gained 10 cents,” Melamed said.

Each of the 16-round games examined in the study included about 25 participants, some of whom participated in multiple games with different scenarios. Some of the games generated random groups of participants, while others comprised participants with existing connections — an arrangement that was designed to mimic real life, where humans often run in packs socially and at work.

Within any given game, players were allowed varying degrees of interaction. In some games, players could interact only with the assigned partners for the duration. In the dynamic networks, participants could cut ties with another player and form new connections.

Furthermore, some of the games included reputation information. Participants were labeled based on their history of willingness to share money. The idea was to test whether those known to collaborate were favored by other players based on reputation — a factor shown in previous research to play a significant role in whether a person is likely to partner with another.

Melamed and his research partners were surprised to find that reputation played no role in collaboration in this study. The findings might have departed from prior studies because of the difference in size and study design, he said, explaining that much of the previous work in this area has been conducted in groups of 100 or fewer and mostly involved student subjects. The participants in Melamed’s study were nationally representative.

Collaboration rates overall were high — and highest when the participants were operating in groups with existing connections and had the ability to drop a partner in favor of another, “which seems intuitive if you think about how we interact in the real world,” Melamed said.

So, if effective teamwork is your goal, your best chance of it is to let people choose their own partners, and switch if/when it doesn’t work out. I sense this could be the beginning of many beautiful partnerships. Dibs on Muffin for chores and Manoj from Accounting for work. That guy is hilarious.