Remote employees reported better well-being than hybrid and onsite colleagues in a new study of 7,704 workers, adding a substantial data point to the argument over return-to-office mandates. The result matters because it tests two familiar claims behind those mandates: that working away from the office harms employee welfare and weakens the connections that keep people at a company. Neither claim held up cleanly in this workplace.
The differences were clear but modest. On a five-point scale, fully remote employees averaged 4.22 for well-being, compared with 4.12 for hybrid workers and 3.89 for onsite staff. The peer-reviewed study in Frontiers in Psychology, published in July, found statistically significant differences between all three groups. Remote workers also used language associated with positive workplace connection slightly more often, although that difference was not statistically significant.
That is useful evidence against treating the office as an automatic cure for disconnection. It is not, however, proof that sending any workforce home will improve its mental health. The research covers one large healthcare organization in the southwestern United States, and different jobs were not randomly assigned to different locations. The strongest reading is narrower: in this organization, remote work coexisted with higher reported well-being without a measurable connection penalty.
What the researchers measured
The researchers used an employee engagement survey conducted in 2023. Of 23,316 permanent employees invited to participate, 20,856 responded, an 89% response rate. The final analysis included the 7,704 people who also answered an open-ended question about organizational culture. The researchers said excluded respondents did not meaningfully differ from the retained group in age, tenure, gender, work arrangement or later turnover.
Managers had already classified each person in the human-resources system as remote, hybrid or onsite according to their scheduled arrangement. The final sample contained 1,869 remote workers, 2,099 hybrid workers and 3,736 onsite workers. Its mean age was 47, and 68% of participants were women. Employees came from across the organization rather than a single department.
Well-being was based on survey responses, while workplace connection was inferred from the words employees chose when asked to describe the culture. The researchers created a dictionary containing terms related to cohesion, inclusion, care and collaboration, then used the LIWC-22 text-analysis software to identify whether each response contained at least one of them. A response was coded either as showing positive connection or not.
This design makes the connection result less direct than a dedicated survey about loneliness, trust or the strength of workplace relationships. It can show that remote employees were no less likely to describe the culture with words such as “team,” “supportive” or “belonging.” It cannot show how often they talked with colleagues or how close those relationships were.
The study also linked the 2023 responses to actual employment records one year later. That is more useful than asking workers whether they intend to quit, since intentions and departures are not the same thing. Turnover was 7.8% among remote workers, 8.3% among hybrid workers and 8.8% among onsite workers. Those location-based differences were not statistically significant.
Well-being, not location, tracked retention
The retention analysis produced a more careful finding than a simple claim that remote work keeps people from quitting. Work arrangement by itself was not a significant direct predictor of turnover. Higher well-being was. In the researchers’ model, each one-point increase in well-being was associated with 29% lower odds of leaving, expressed as an odds ratio of 0.71.
The paper found a statistically significant indirect relationship in which remote or hybrid arrangements were associated with better well-being, which was then associated with lower turnover. The estimated effect was minimal, and the total effects comparing the work-location groups were not significant. In plain language, happier employees were less likely to leave, and remote employees happened to report greater well-being, but the records do not show that location alone kept them at the company.
That distinction is easy to lose in a debate often reduced to remote versus office. The findings point instead toward employee experience as the meaningful variable. A mandate can change where people sit without addressing workload, management quality, scheduling or whether a team communicates well. Likewise, remote status alone cannot repair a poorly run workplace.
Stefanie Johnson, a University of Colorado Boulder professor and study co-author, told CU Boulder Today that employers should preserve remote work when employees want it. She pointed to control over the work environment and the removal of commuting and care-related logistics as possible explanations for higher well-being. The study itself did not test those mechanisms, so they remain informed hypotheses rather than findings from this dataset.
The office did not guarantee connection
Remote employees were slightly more likely to use a positive-connection term in their culture descriptions than hybrid or onsite workers. But because the gap was not statistically significant, the paper cannot establish that remote workers were more connected. It can reasonably say that this sample showed no evidence of the expected connection deficit.
That matters for technical organizations, where distributed work is often supported by issue trackers, code review, shared documents, chat and video. Physical proximity is only one input into collaboration. Response norms, documentation quality, meeting design and access to decisions can matter just as much. A worker in the same building can still be excluded from information; a distributed team can make access explicit.
The authors argue that communication quality may explain why distance did not correspond with weaker connection. They recommend clear team norms covering response expectations, meeting etiquette, conflict and check-ins, with regular reviews as roles and work arrangements change. These are practical suggestions, but the study did not compare particular tools or communication routines. It therefore cannot identify which remote practices produced the observed result.
There are also situations in which being together is plainly useful. Johnson noted that face-to-face time can help people build relationships, particularly early in a career or before coworkers know one another. The study does not test onboarding, mentoring, creative work or individual team performance. Its finding is not that offices serve no purpose. It is that mere attendance should not be used as a stand-in for culture or connection.
Why the result cannot settle the debate
The largest limitation is selection. Clinical and operational healthcare roles are much more likely to require onsite work, while administrative and analytical roles are more likely to be remote. Those jobs can differ in physical demands, emotional strain, schedules, pay, autonomy and exposure to patients. The researchers acknowledge that unmeasured job characteristics could explain part of the well-being gap.
Workers also were not randomly assigned to locations. People who prefer remote work may seek remote-compatible roles, and managers may grant flexibility to particular jobs or employees. That makes it impossible to conclude from this observational study that changing an onsite employee’s location would raise their well-being by the reported amount.
Generalisability is another constraint. Studying one organization controls for a shared culture and set of systems, but the same feature limits application to other employers, industries and countries. A healthcare institution with established remote practices may differ sharply from a young software company, a public agency or a manufacturer. The underlying data are not public because of privacy and ethical restrictions, although the paper says de-identified data may be available from the corresponding author under appropriate conditions.
Finally, well-being was self-reported, and the connection measure reduced short culture descriptions to a yes-or-no keyword signal. The authors call for longitudinal or experimental research, controls for job type and objective indicators such as absence or health claims. Those additions would help separate the effect of location from the characteristics of the work and the people doing it.
What employers can take from it
The study does not deliver a universal schedule. It does raise the evidentiary bar for a universal mandate. An employer arguing that attendance will improve welfare, connection or retention should define the problem, measure it and test whether location is actually the cause. This dataset shows that an office-heavy workforce can report lower well-being without showing stronger cultural connection or meaningfully lower turnover.
For employees and managers, the practical lesson is similarly restrained: flexibility is a plausible contributor to well-being, but job design and communication still shape the result. Organizations considering policy changes should compare teams and roles, track outcomes before and after changes, and distinguish collaboration needs from a preference for visible attendance.
What to watch next is whether studies across multiple organizations can reproduce the well-being gradient while controlling for occupation, autonomy and employee preference. Randomized or natural experiments would be especially valuable, as would direct measures of loneliness, mentoring and team performance. Until then, this research is strong evidence that remote work need not damage connection, not a final answer about where every employee should work.