A 2026 study of 7,704 employees found no detectable difference in how connected remote, hybrid, and onsite workers felt to remote work and company culture. Remote employees reported the highest well-being and the lowest turnover, though the culture gap was too small to distinguish from chance.
The honest summary is narrower and more useful than the headlines: the research found no evidence that distance damages workplace connection. It did not find that remote work improves culture. Those are different claims, and conflating them weakens the argument.
Here is what the study did, what it found, what it cannot tell you, and what follows for building distributed engineering teams.
What the study actually was
| Element | Detail |
| Title | Rethinking return-to-office: the positive impact of remote work on well-being, connection, and employee retention |
| Published | July 2026, Frontiers in Psychology |
| Sample | 7,704 employees at the University of Texas MD Anderson Cancer Center |
| Co-author | Stefanie Johnson, professor of organizational leadership and information analytics, University of Colorado Leeds School of Business |
| Design | Time-lagged, survey completed in 2023 and linked to employment records one year later |
| Groups | 1,869 fully remote, 2,099 hybrid, 3,736 fully onsite |
Why the design is stronger than most remote work research: it is time-lagged rather than cross-sectional. Researchers measured attitudes at one point and then checked actual employment records a year later, so the turnover finding reflects what people did rather than what they said they might do.
The University of Colorado Boulder’s announcement and Fortune’s coverage both carry the researchers’ own framing of the findings.
The three findings, with their actual strength
| Finding | Result | How strong |
| Well-being | Remote 4.22, hybrid 4.12, onsite 3.89 on a five-point scale | The strongest finding |
| Culture and connection | Remote 61%, hybrid 58%, onsite 55% used positive relational language | Too uncertain to rule out chance |
| Turnover after one year | Remote 7.8%, hybrid 8.3%, onsite 8.8% departed | Real but small |
On well-being: the measure was not a single happiness question. It combined four statements covering whether employees felt the organisation cared about their health, whether they felt energised by their work, whether their manager supported their well-being decisions, and whether they could maintain balance in their lives.
On culture: employees were asked to describe their organisation’s culture in a handful of words, and researchers looked for terms associated with care, inclusion, teamwork, and belonging. Remote workers used that language slightly more often. As the detailed breakdown shows, once the researchers accounted for the different group sizes, the gap could not be distinguished from chance.
On turnover: a one percentage point spread across a year. Real, consistent in direction, and modest.
The finding everyone is misreporting
This is the part worth getting right, because citing it carelessly will cost you credibility with anyone who reads the paper.
What the study shows: an absence of evidence that remote work harms connection.
What the study does not show: evidence that remote work improves connection.
Those sound similar and are not. The first is a null result, which is genuinely informative here because it directly contradicts the central premise of most return-to-office mandates. The second is a positive claim the data does not support.
Why the distinction matters practically: if you argue internally that remote work improves culture and someone reads the paper, you lose the argument and the credibility. If you argue that the assumption behind the mandate is unsupported, you are on solid ground.
The defensible version of the claim: in this population, distance did not produce the disconnection that return-to-office arguments assume it produces.
Six limitations to know before you cite this
Anyone using this study should be able to name its weaknesses first.
1. Single organisation, single sector: All 7,704 employees worked at one healthcare organisation in the southwestern United States. Healthcare has different role structures, collaboration patterns, and career norms than software engineering. Generalising directly to an engineering team is a stretch.
2. Self-selection: People who wanted to work remotely were likely more able to do so, and the roles that permit full remote work differ systematically from those that do not. The comparison is not between identical people in different arrangements.
3. Choice may be the actual variable: The researchers themselves emphasise flexibility over any particular arrangement. Johnson’s framing is that removing people’s choice about how they work is unlikely to help their well-being, which suggests autonomy rather than location is doing much of the work.
4. No productivity measure: The study did not track output. Anyone citing it as evidence that remote work is more productive is citing something it does not contain.
5. Timing: The survey was completed in 2023 and published in 2026. Working patterns, tooling, and organisational norms have moved since.
6. The culture result is a null, not a positive: Covered above, and the most common misreading.
What the study does support
Strip out the overreach and a clear, defensible position remains.
1. The central argument for return to office is unsupported in this data: The claim that people must be physically together to feel connected did not hold in a sample of nearly 8,000 employees with a year of follow-up employment records.
2. Flexibility appears to matter more than arrangement: The consistent pattern across well-being and retention points toward autonomy rather than toward any specific location policy.
3. Face-to-face time still has a role, particularly early in careers: Johnson’s own caveat is direct on this point, noting that remote arrangements work better once people already know each other. The implication is that some in-person contact retains value, especially for people building their first professional relationships.
4. If you require office attendance, make the time worth attending for: The researchers’ practical suggestion is that when people do come in, there should be a reason that makes the time valuable rather than an expectation that presence alone produces connection.
What this means for engineering teams specifically
The study population was healthcare, so translation requires care. Three things carry across reasonably well.
Connection is built by structure, not proximity: If distance does not automatically produce disconnection, then whatever does produce it is something else, which is consistent with what distributed engineering teams report: documentation, decision visibility, and communication norms determine cohesion more than location does.
Choice is a retention lever: The one percentage point turnover difference is small, but the direction is consistent and the mechanism, autonomy, is well established in the broader literature. For engineering roles where replacement cost runs into six figures, a small retention effect is not trivial.
Early-career people need more deliberate attention: The face-to-face caveat maps onto a real engineering problem, which is that junior engineers build judgment through supervised practice and informal observation. Distributed teams have to construct that deliberately rather than assume it.
What does not translate: Healthcare collaboration patterns, shift structures, and the physical presence requirements of clinical work are not comparable to software delivery. Do not cite the specific numbers as though they describe an engineering population.
What the study does not settle
Honesty about the open questions is what makes the rest credible.
| Open question | Why this study cannot answer it |
| Does remote work affect productivity? | Productivity was not measured |
| Does it affect innovation or idea generation? | Not measured |
| What happens to junior development over years? | One year of follow-up only |
| Does this hold in engineering organisations? | Single healthcare organisation |
| Is location or choice the causal factor? | Design cannot fully separate them |
| What about fully distributed international teams? | Sample was one US organisation |
The broader evidence picture remains mixed. Earlier survey research found employees reporting that specific aspects of culture had suffered under remote working, particularly team spirit, colleague support, and learning objectives. The honest position is that the literature does not converge on a single answer, and anyone claiming it does is selecting their sources.
What to do with this
Four practical conclusions, in order of how well the evidence supports them.
Stop treating disconnection as automatic: The assumption that distance causes disconnection is not supported here, which means if your distributed team feels fragmented, the cause is probably something you can name and fix rather than the arrangement itself.
Preserve choice where the role allows it: The strongest signal in the data points at autonomy. Where a role can be done remotely, removing that option has a measurable cost in employee well-being and a smaller one in retention.
Build connection deliberately rather than architecturally: Documentation, decision records, communication norms, and rotated meeting times do more for cohesion than an office does. That is consistent with both this study and practitioner experience.
Invest in-person time where it has the most effect: Early-career people, new joiners, and the start of a team’s life are where face-to-face contact appears to matter most. Spending the budget there beats spreading it evenly.
The hiring consequence
If location does not determine connection, then location should not determine your hiring pool either. That is the practical implication for anyone still restricting recruitment to a commutable radius.
The constraint that remains is access rather than geography. Senior engineers in scarce specialisms are employed, performing well, and not applying anywhere regardless of where they live, which is why widening the map does not automatically widen the pool. That is the case for off-market sourcing through closed networks.
Frequently asked questions
No. The study found no detectable difference in how connected remote, hybrid, and onsite employees felt to their workplace culture. Remote workers used slightly more positive relational language, but after accounting for group sizes that gap was too uncertain to rule out chance, so the accurate statement is that distance did not damage connection rather than that it improved it.
A year after the survey, 7.8% of remote employees had left, compared with 8.3% of hybrid and 8.8% of onsite workers. The direction is consistent but the spread is about one percentage point, so it is a real effect rather than a dramatic one. For expensive-to-replace roles, even a small retention difference has meaningful cost implications, and the full employment cost comparison sits in transparent pricing.
The study suggests choice matters more than arrangement, which implies the answer varies by person. One of the researchers noted that remote arrangements work better once people already know each other, which points to a genuine limitation for new joiners and early-career employees. Practitioner accounts consistently describe outcomes as bimodal, with some people thriving remotely while others struggle with isolation and loss of routine.
Partially. The mechanisms, autonomy and deliberately built connection, translate well, but the specific numbers come from healthcare workers with different collaboration patterns and role structures. Use the study to challenge the disconnection assumption rather than to predict outcomes for an engineering population.
What to take from this
The choice is narrower than the marketing suggests. Establish whether the person is an employee or a genuine contractor, establish whether you own a local entity, and five of the six options usually eliminate themselves.
Then model total cost per country rather than comparing provider fees, because statutory variation dwarfs fee variation.
And keep the sourcing decision separate from the employment decision, because conflating them is how companies end up paying an ongoing percentage for something a one-time arrangement would have covered.
To work through which structure fits your situation, get in touch.
The honest conclusion
The study does not prove remote work is better for culture. It shows that in one large organisation, over a year, the disconnection that return to office arguments assume simply did not appear.
That is a narrower claim than the headlines, and it is the one worth making, because it is the one that survives scrutiny.
The more interesting finding sits underneath: the consistent pattern across well-being and retention points at choice rather than location. Which suggests the useful question is not whether people are in a building, but whether they had any say in it.
To talk through building a team where connection is designed rather than assumed, get in touch.
