Coworking research: productivity, loneliness, community

The short answer, and why it is not simple

If you are deciding whether a coworking membership is worth paying for, the research gives you less certainty than the marketing does. Here is the evidence in brief:

• Productivity. The strongest studies are randomized experiments, and they compare home with the employer's office. None of them tests a coworking space. Their results also point in different directions depending on the job.
• Retention and satisfaction under hybrid work. The evidence is strong, but it comes from one large firm in China.
• Coworking and social life. The research is real but thin: one qualitative case study, a few cross-sectional surveys and a magazine-published survey. It suggests that coworking spaces can provide community and social support. It does not show that joining one causes you to feel less lonely or to produce more.
• Loneliness and work. Loneliness is linked to lower job performance in field data. Reviews of telework say the literature is fragmented and mostly cross-sectional.

The rest of this article goes through each study: what it measured, on whom, how large the effect was, and what it cannot tell you.

How to judge a workplace study

Four questions sort most workplace studies into "strong" and "suggestive":

Randomization matters most. People who choose coworking differ from people who don't: in income, personality, type of work and need for company. A survey that finds coworkers happier cannot separate the effect of the space from the type of person who signs up. Peer review is a weaker filter, but a real one. The NBER, for example, states on its working papers that they "have not been peer-reviewed" ^[1].

What coworking research exists

Research on coworking itself falls into three groups:

1. Qualitative case studies. Researchers observe and interview members of one or a few spaces. These explain how community forms. They cannot measure how much it helps or generalize to other spaces.
2. Cross-sectional surveys. Members answer a questionnaire once. These show associations (for example, social support goes with higher satisfaction) but not cause and effect.
3. Industry reports. Operators, brokers and landlords publish them, often from their own members. They are useful for describing the market, but they have a commercial interest in the result. This article does not rely on them.

We found no randomized trial that assigns people to a coworking space and measures their output or loneliness against a home or office group.

Garrett, Spreitzer and Bacevice (2017): how community gets built

Study: Lyndon Garrett, Gretchen Spreitzer and Peter Bacevice, "Co-constructing a Sense of Community at Work: The Emergence of Community in Coworking Spaces," Organization Studies 38(6), 821–842 ^[2].

Design: The abstract describes it as "a qualitative, single case study" of one coworking space. The authors define coworking spaces as places where people "do their own work but in the presence of others with the express purpose of being part of a community."

Finding: Members build a sense of community through day-to-day interactions. The authors name three types of collective action that do this: endorsing, encountering and engaging. Their main point is that community is something members do, not a feature the space supplies.

Limits: It is one space. A single case study can build theory, but it cannot say how common these behaviours are or whether they reduce loneliness. It also says nothing about productivity. In practice, the finding cuts both ways: a space where members don't take part in community work may feel like a quiet open-plan office, whatever its brochure says.

Gerdenitsch and colleagues (2016): social support, measured

Study: Cornelia Gerdenitsch, Tabea Scheel, Julia Andorfer and Christian Korunka, "Coworking Spaces: A Source of Social Support for Independent Professionals," Frontiers in Psychology 7 ^[3].

Design: Two studies. Study 1 surveyed 69 coworkers and found that social interaction in coworking spaces can take the form of social support. Study 2 surveyed 154 coworkers and compared them with 609 employees in traditional organizations.

Finding: Social support from both sources was positively related to performance satisfaction, meaning how satisfied people were with their own performance. Among coworkers, the self-efficacy pathway appeared only when time pressure was high. The authors conclude that "a mobilization of social support seems necessary in coworking spaces." In other words, support in a coworking space has to be sought out; it cannot be assumed.

Limits (the authors' own): Participants were not randomly selected, so more integrated coworkers may have been more willing to respond. Study 2 is cross-sectional, which "does not allow for causal inferences." The data are self-reported. The outcome was satisfaction with performance, not measured output.

Survey evidence on social fulfilment (Hadley and colleagues, 2023)

Study: Constance Noonan Hadley, Ben Marks and Sarah Wright, "Research: How Coworking Spaces Impact Employee Well-Being," Harvard Business Review digital article, February 2023 ^[4].

Design and sample: A survey of 819 employees in office roles, across countries and industries ^[5].

Finding: Respondents rated "third spaces" (cafés, hotel lobbies, friends' houses and coworking sites) as more socially fulfilling than working from the office or from home. The authors link the result to "relational job crafting": people can choose whom they interact with, avoid unwanted coworkers and escape the evaluative pressure of their own office.

Limits: The survey is self-reported and cross-sectional. It was published in a practitioner magazine, not a peer-reviewed journal. Third spaces are grouped together, so the result is not specific to coworking. It also covers employees of organizations, not freelancers. It is the most directly relevant survey for a hybrid employee, and still the weakest design in this article.

Productivity: what the Ctrip experiment says about working from home

Study: Nicholas Bloom, James Liang, John Roberts and Zhichun Jenny Ying, "Does Working from Home Work? Evidence from a Chinese Experiment," Quarterly Journal of Economics 130(1), 165–218 ^[6]. An earlier version is NBER Working Paper 18871.

Design: Ctrip was a 16,000-employee travel agency. Call-centre employees who volunteered to work from home were randomly assigned to home or office for nine months.

Finding: Performance rose 13%. About 9 points came from working more minutes per shift (fewer breaks and sick days) and 4 points from more calls per minute, which the authors attribute to a quieter, more convenient environment. Home workers reported higher work satisfaction and their attrition rate halved. Their promotion rate, conditional on performance, fell. When the firm later let everyone choose, over half of the experimental employees switched, and the measured gain from working at home nearly doubled to 22%.

What it cannot tell you about coworking: The comparison is home against a busy call-centre office, for individual work that is easy to count. A coworking space is neither of these. The finding that quiet raised output is a reminder that a noisy shared space can cost productivity as easily as it adds it. The sample was also volunteers, and the NBER version discloses that co-author James Liang was Ctrip's chief executive at the time.

A contrary result: data entry in India

David Atkin, Antoinette Schoar and Sumit Shinde ran an RCT in the data entry sector in India, assigning workers to home or office ^[7]. Workers randomly assigned to work from home were 18% less productive than those in the office. Two-thirds of the gap appeared from the first day, and the rest came from faster learning in the office. Workers who preferred home-based work were less productive at home than in the office: 27% less, against 13% for workers who preferred the office. The authors partly explain this by home demands such as childcare. This is a working paper, not a peer-reviewed article.

Another: IT professionals during the pandemic

Michael Gibbs, Friederike Mengel and Christoph Siemroth studied personnel and analytics data from over 10,000 professionals at an Indian technology company, before and during pandemic home working ^[8]. Hours worked went up and output declined slightly, so productivity fell 8% to 19%. Time on meetings and coordination increased, uninterrupted work time shrank, and employees networked with fewer people. This is observational data from a forced, pandemic-era switch, not a randomized trial.

Taken together, the effect of working from home on productivity depends on the task, the home setting and who chooses it. None of these studies measures a coworking space.

Hybrid work and retention: the 2024 Nature trial

Study: Nicholas Bloom, Ruobing Han and James Liang, "Hybrid working from home improves retention without damaging performance," Nature 630, 920–925 (2024) ^[9].

Design: A six-month RCT from August 2021 to January 2022 with 1,612 graduate employees (engineering, marketing and finance) in the Airfare and IT divisions of Trip.com. Employees with odd-numbered birthdays got the option to work from home on Wednesday and Friday. The others came in five days a week.

Finding: Hybrid work improved job satisfaction and cut quit rates by one-third, from 7.2% to about 4.8%. The reduction was significant for non-managers, women and people with long commutes. Equivalence tests showed no effect on performance grades over two years of reviews, and the authors found no evidence of a difference in promotions. The 395 managers' surveyed estimate of hybrid work's effect on productivity moved from −2.6% before the trial to +1.0% after it.

Limits: It is one firm in one country, and the employees were graduates. The competing-interests statement says no funding came from Trip.com, but co-author James Liang is the company's co-founder and chairman with equity holdings. The trial was registered after it began. For coworking, the relevant lesson is indirect: two days away from the office did not hurt measured performance. Where those two days are spent is a separate question the trial does not answer.

Loneliness and remote work: what reviews conclude

The case for coworking as a cure for loneliness rests on two links: remote work can be lonely, and loneliness hurts. The second link has better evidence than the first.

• Loneliness and performance. Hakan Ozcelik and Sigal Barsade ran a time-lagged field study of 672 employees and their 114 supervisors in two organizations. Greater workplace loneliness was related to lower job performance, through lower approachability and weaker commitment to the organization ^[10]. This is correlational. The time-lagged design and the inclusion of supervisors make it stronger than a one-time self-report survey.
• Loneliness in telework. Elisabeth Figueiredo, Clara Margaça and José Carlos Sánchez-García reviewed 65 peer-reviewed articles in a PRISMA-guided bibliometric review ^[11]. They report that the absence of physical interaction and structured social environments worsens feelings of detachment, which contributes to emotional strain and reduced job performance. They also say the literature "remains fragmented." In their limitations, they note that most studies appeared after 2020, many used cross-sectional designs "which restrict causal inference," and loneliness and performance are measured inconsistently. Their first recommendation for future research is longitudinal studies.

Neither source tests coworking as a remedy. The plausible chain is "remote work → isolation → lower well-being and performance → coworking restores contact," but no study has tested the last step.

Confounders and selection effects

Three biases run through this literature:

1. Self-selection into coworking. Coworking members chose to pay for it. They may be more sociable, better funded or more motivated before they arrive. Gerdenitsch and colleagues flag a related problem among survey respondents: integrated members are more likely to answer.
2. Volunteers in experiments. The Ctrip participants volunteered for home working. Atkin and colleagues show that preferences predict productivity in non-obvious directions. A result for volunteers may not hold for the people who would rather stay put.
3. Interested parties. Operator-funded surveys have an obvious stake in the result. Academic work is not immune either: both Ctrip/Trip.com studies include a firm insider as co-author, which the papers disclose.

When you read a headline such as "coworkers are X% more productive," ask whether the people were randomized, whether productivity was measured or self-rated, and who paid for the survey.

What the evidence suggests for different people

For a Montreal hybrid worker, the commute effect in the Nature trial is relevant: the biggest drop in quits was among people with long commutes. A space near home may keep the commute savings while adding company. Our comparison of coworking and a home office covers the practical trade-offs, and this analysis of remote and hybrid job offers prices what those home days really cost.

A way to test coworking for yourself

Since nobody has run the randomized trial for coworking, you can run a small one on yourself. It will not be scientific, but it beats guessing.

1. Pick one output measure before you start. Use something you already count: billable hours, words drafted, tickets closed or sales calls made. Don't use "felt productive."
2. Add two daily ratings, from 1 to 5, at the end of each day: how connected you felt to other people, and how well you concentrated.
3. Alternate locations for four weeks. For example, spend weeks 1 and 3 at home and weeks 2 and 4 at a coworking space on the same weekdays. Alternating spreads out deadlines and seasonal swings.
4. Use day passes, not a contract. The day pass versus monthly membership guide shows where passes stop being the cheaper option.
5. Compare averages, then discount the novelty. A new place often feels better in week one. If the difference holds in week 4, it is more believable.
6. Price it. Divide the monthly cost by the extra output, or by the days you rated as "connected." Then decide whether that is worth it to you.

Bottom line

• Working from home affects productivity in both directions. The size and sign depend on the task, the home and who chooses it, as the Ctrip, Indian data-entry and IT studies show.
• Hybrid work kept people longer without hurting their reviews in one large, well-designed trial.
• Coworking's social benefits are plausible and supported by small studies and surveys, but no study shows that coworking causes lower loneliness or higher output.
• Missing research: randomized or longitudinal studies that compare coworking with home and office on measured output and validated loneliness scales, outside a single firm or space.

Until those studies exist, the honest case for coworking is personal: try it, measure it and keep it only if your own numbers support it.

If you want to run that test in Griffintown, 2727 Coworking sells day passes from $55 before taxes (as of October 2026) at 2727 Rue Saint-Patrick, on the Lachine Canal.