Full Remote Work Isn't for Everyone — Reading the 'Home Alone' Isolation Study from Both Sides of the Household

Full remote work is comfortable. No commute, your own rhythm, more time with family. Having worked remotely for a long time, I know those benefits in my bones.

But what if, behind that comfort, a different cost is quietly accumulating? The study “Home alone: Remote work, isolation, and mental health,” published in Science on June 4, 2026, is precisely an attempt to quantify that hidden side, using data from nearly 590,000 workers.

I have done full remote work both while living alone and, now, while living with my wife and three children in a household of five. The study’s single most important finding is that the harm concentrates among people who live alone, and when I think back across my two living situations, that conclusion lands almost painfully well. In this post I read the study’s data alongside my own experience.

What the study found — working from home worsens isolation and mental health

Let me start with the research itself. The authors are Natalia Emanuel (Federal Reserve Bank of New York), Emma Harrington (University of Virginia), and Amanda Pallais (Harvard University). They use five nationally representative U.S. surveys (2011–2024, N = 588,322), excluding the peak pandemic years of 2020–2021.

The clever part is the design. The COVID-19 pandemic spiked remote work specifically in “remotable” jobs such as software engineering and marketing, not in jobs that require physical presence like manufacturing or nursing. The authors split occupations along that line using the Dingel–Neiman remotability index, treat remotable jobs as the treatment group and on-site jobs as the control, and apply a difference-in-differences approach to isolate the effect of remote work itself. Because they study occupation-level shifts rather than individual choices, they sidestep the reverse-causation worry that “maybe people who were already struggling simply chose to work from home.”

Here is what the data showed.

Measure (pre/post difference-in-differences)	Change for remotable jobs
Share of fully remote workdays	+17.9 pp (differential vs. control)
Time spent alone per workday	about +1 hour (+58.0%)
Days spent entirely alone	84.0% in 2022–2024 (vs. 23.2% on-site)
K-6 psychological distress score	+0.1 standard deviation

As remote work surged, people in remotable jobs spent roughly one additional hour alone per workday, and that time shifted away from collaboration and toward solo work. After-hours socializing with friends declined too. The K-6 (Kessler Psychological Distress Scale), a measure of mental distress, also deteriorated.

The authors estimate that remote work explains roughly one-third (32–36%) of the overall rise in isolation and mental distress in society between 2011–2019 and 2022–2024.

The key finding — the harm concentrates among those who live alone

What makes this study compelling reading is that it digs past the average to ask who actually gets hurt.

The worsening of isolation was sharply concentrated among people living alone. The increase in days spent entirely alone was about 10 times as large for those living alone as for those living with others (7.0 pp vs. 0.7 pp). And the increase in days with no human contact at all (no chitchat with a barista, no nod to a cashier, not even a passerby on the street) was about 13 times as large (3.9 pp vs. 0.3 pp).

Mental health follows the same pattern. The rise in psychological distress was about twice as large for people living alone as for those living with family. The starkest evidence is in the prescription data: among remote workers living alone, prescriptions for depression and anxiety rose by a differential 5.1 pp, and use of mental-health medications overall rose by 5.3 pp. They were also 4.6 pp more likely to see a mental-health professional.

For people who live with others, these effects were either statistically insignificant or far smaller. In other words, it isn’t “working from home” as such; it’s the combination of “working from home and living alone” that pulls the trigger on isolation and declining mental health. That combination is the study’s central finding.

When I was living alone and fully remote

Now for my own story. There was a period when I lived alone and worked fully remotely, exactly the environment in which the study’s “days spent entirely alone” can quietly materialize if you let them. A day where I never stepped outside and only used my voice for work video calls was, honestly, easy to produce.

But back then I consciously avoided that. I made a deliberate effort to see people outside of work hours.

I was living in a big city at the time, which made it easy to stay mobile. I would find an interesting meetup on a community event platform, and if my schedule allowed, show up in person. I’d chat with strangers over some technical topic and head home. Looking back, that was a hobby, but it was also, in the study’s vocabulary, an active way of killing off the “day spent entirely alone.” I was recovering, outside of work, the socializing that remote work had stripped from my workday.

Even on days when I didn’t interact with anyone directly, one small habit helped. As I wrote in an earlier post, it was listening to podcasts. Just having a human voice in the room settled me, oddly enough. The main goal was to absorb knowledge from tech podcasts, but a side effect was that it softened the state of being “alone in complete silence.” Reading the study’s emphasis on the harm of “days with no human contact,” I think I finally understood, in hindsight, what that unconscious defense had been doing for me.

Now in a household of five — same remote work, different beast

These days I live in a household of five: my wife and our three kids. Even though it’s the same full remote work, the situation I’m in looks completely different through the study’s lens.

The study’s finding that “the effect is far smaller for those living with others” maps exactly onto my lived experience. When I finish work and step into the living room, there are people there whether I like it or not. A child wants to talk, I’m called to help with dinner, someone wants help with homework. Spending an entire day completely alone is structurally impossible. The study finds that days spent entirely alone rose tenfold for people living alone, but in a household of five, that kind of day simply never arrives.

You could say a family-based household quietly underwrites the comfort of full remote work. Which also means: for someone living alone, the safety net I’m enjoying doesn’t exist. The same arrangement goes by the same name, yet the cost each person pays is utterly different, and the study makes that point in hard numbers.

Remote work isn’t for everyone — a friend’s choice

There’s one more example I keep coming back to: a friend of mine (not a colleague).

Full remote work just didn’t suit him, and he changed jobs to a company that mandates one day in the office per week. It wasn’t a skills problem or a performance problem. As he kept working alone at home all day, his mental health actually broke down: a persistent low mood set in, and he felt he couldn’t keep going under that arrangement. He functions better when there’s a place to physically see coworkers, even once a week. For that reason, he deliberately switched to a job with more constraints on how he works.

At the time I filed it away as “some people are like that,” but having read this study, his choice now looks entirely rational. Among its policy implications, the paper highlights the importance of designs that reduce isolation: coordinating in-office days for hybrid teams, encouraging informal interaction even online. Without leaning on any company policy, my friend chose his own design of “one office day a week.” I think he understood the cost structure of his own mental health more accurately than anyone.

Full remote work isn’t for everyone. This isn’t a matter of willpower; it’s that fitness for it splits cleanly along fairly blunt conditions: your living situation (alone or with others) and your temperament (whether time alone wears on you).

What the study might be missing — compensatory behavior

That said, the study has an important limitation that the authors themselves candidly acknowledge, and against my own experience, it’s the point that resonates most.

One is that they can’t distinguish fully remote from hybrid work. Given the data, the two aren’t separated, so the authors note that “one or two days of office presence per week may have considerably smaller impacts, or even be protective.” The hypothesis that my friend’s “one office day a week” works has not, at the level of this study, been tested yet.

The other, decisive for me, is that the data ends in 2024. The authors write: “If those in remotable jobs have made compensatory changes (such as cultivating social connections outside of work), they may not yet have reaped the full benefits.”

That is exactly what I was doing back in my living-alone years. Going to meetups, putting a human voice in the room with podcasts. It was compensatory behavior: deliberately re-filling, outside of work, the socializing that remote work had taken from the job. The study’s data may be a snapshot of a transitional moment, captured before such adaptations had a chance to spread.

Turned around, this means the cost of full remote work changes a great deal depending on whether you take compensatory action. The deterioration the study describes should be read not as “the inevitable fate of working from home” but as “the default value when no countermeasures are taken.” Even living alone, if you build a mechanism into your life that actively kills off isolation, you can lower the cost. And conversely, even someone with a structural safety net, like my household of five, has other pitfalls if they grow complacent.

Conclusion — know your own cost structure

I’ve read Science’s “Home alone” against my own two stints of full remote work: living alone and living in a household of five. The three things I took away:

Working from home worsens isolation and mental health. Remote work explains roughly one-third of society’s overall rise in isolation and distress, an effect large enough to show up even in the average.
The harm concentrates among those who live alone. Days spent entirely alone rose about 10x more than for people with housemates, and days with no human contact about 13x more. The same arrangement carries an order-of-magnitude different cost depending on your living situation.
Compensatory behavior lowers the cost. Cultivating social connections outside work, keeping a human voice in your life: these help. The study’s data is merely a snapshot from before such habits had spread.

Full remote work isn’t for everyone. But that’s not the same as telling unsuited people to give up. Once you understand your own cost structure (whether you live alone or with others, whether time alone wears on you), you have plenty of options: build the necessary compensatory behavior into your life, or, like my friend, redesign the way you work itself.

The cost that quietly accumulates behind comfort is hard to notice until someone puts a number on it. The greatest value of this study is that it makes that cost visible and gives you a reason to ask: which side am I on?

That’s a view from the trenches of someone who has done full remote work both alone and in a household of five, reflecting on a study of isolation and mental health.