Picture a lit-up carriage on the last train home. Every seat is taken, the window is full of other people’s reflections, and somebody’s laugh carries down the aisle. You are in the middle of all of it. Your phone is warm in your hand, your coat is brushing a stranger’s sleeve, and you have never felt further from anyone. Nobody has done anything wrong. You are simply in a crowd and, at the same time, on your own.
If that carriage is a place you recognise, you may have been told, kindly or otherwise, that you cannot possibly be lonely, because look at all the people in your life. It might be the friend who texted at lunchtime to ask how you are, and the “fine, you?” you sent back. It might be a parent holding a letter from school, or someone waiting for an ADHD assessment, or someone whose diagnosis is only a few weeks old, with plenty of people around and nobody who quite gets it. A large new body of research suggests that those well-meaning remarks have been answering the wrong question.
This post looks at what that research found, what it cannot tell you, and what you can try this week.
Two measurements that are not the same thing
Researchers have long separated two things that sound identical and are not. Social isolation is objective: how many people you see, how often, and how many you could call. You can count it. Loneliness is subjective: the felt gap between the closeness you have and the closeness you want. Nobody can count that for you, and nobody can talk you out of it with a headcount.
In 2026, Ong and colleagues published a study in JAMA Network Open that put the two side by side. They followed 7,845 adults over the age of 50 in England for an average of 13.6 years. Rather than asking only whether people were isolated, or only whether they were lonely, they looked at the gap between the two. They called it social asymmetry: how far someone’s loneliness runs ahead of, or behind, what their circumstances would predict.
People who were lonelier than their circumstances predicted, whom the researchers called socially vulnerable, had an elevated risk of dying from any cause, of cardiovascular disease and of chronic obstructive pulmonary disease (COPD, a long-term lung condition). People who were objectively isolated but not lonely, the socially resilient, had little increased risk. The larger the gap between how lonely someone felt and how connected they looked on paper, the greater the cardiovascular and mortality risk. That held after the researchers adjusted for demographic, behavioural and health factors.
The study was reported by the Cornell Chronicle in March 2026, and the message Ong took from it was that connection alone is not the whole story.
Read that again on the train, or at half past eleven on a bad night with your thumb hovering over a message you have not answered. It says something specific to you. The feeling is not a flaw in your character. It is something researchers now measure in its own right.
What this study does and does not say
I want to slow down here, because hope without honesty helps no one. These limits are not a reason to dismiss the findings. They are there to protect you, so that you do not carry a headline to your GP, or to the friend who keeps telling you that you are fine, and expect it to bear more weight than it can.
The study is observational. It shows that a pattern and a health risk travel together, but it cannot prove that loneliness caused the outcomes. Everyone in it was over 50 and living in England, so if you are twenty-six, or a parent of toddlers in Leeds, the findings are suggestive rather than directly about you. The effects were modest for any one person, and significant at population scale, which is a different claim from “this will happen to you”. A raised risk is not a diagnosis. It is not a sentence either.
And a study cannot see your evenings. Only a conversation with someone who can ask you questions can do that.
What the research can do is change the question. If you have been asking “why am I like this, when I have so many people?”, it suggests a kinder and more useful one: what kind of closeness is missing?
Why a full diary and an empty feeling can sit side by side
A companion study, published in Communications Psychology, helps to explain how. It was small and close-up. It followed 157 adults for 20 days, prompting them on their phones five times a day. This method is called experience sampling, and it catches feelings as they happen instead of asking people to remember them weeks later.
Loneliness, it turned out, was not a fixed trait that some people simply have. It moved. The thing that most strongly predicted a lonely moment was perceived social threat: the sense of being excluded, judged or devalued, whether or not anyone had actually done any of it. The researchers also found a self-reinforcing cycle, in which loneliness and the sense of threat feed one another. Ong described it as a flywheel: once it is turning, each moment adds momentum.
Now picture it in the ordinary detail of your own week. A colleague says “we should catch up properly sometime”, and you hear the word “properly” as a polite way of saying not now. A friend’s reply is warm and brisk, and you read it as slightly cooler than usual. At dinner you decide not to mention the hard thing, because they have had a long day. Each decision is small and sensible. Added together, they are how you end a good-looking Friday night feeling as though you spent it behind glass.
Have you noticed that your own loneliness has a time of day, or a type of message that sets it off? That is worth knowing, because a flywheel can be nudged at any point on its turn.
In that study, people with chronic loneliness showed stronger and more persistent links between these moments. This is not a verdict on you. It simply suggests that a system which has been turning for a long time has more momentum.
It also explains why advice that works for a neighbour can leave you cold. If perceived threat is part of what keeps the wheel turning, then being handed more social occasions without any change in how safe they feel can simply give the threat more places to show up. Safety, in this sense, is not about the room being quiet. It is about whether you expect to be met kindly when you say something true.
Social loneliness and emotional loneliness are different hungers
The distinction goes back decades in the loneliness literature, usually credited to the sociologist Robert Weiss. Social loneliness is the shortage of a wider network: people to do things with, a community, a place in a group. Emotional loneliness is the shortage of someone who really knows you, someone you can say the unedited version to. You can be rich in the first and poor in the second, which is exactly the experience of being surrounded and alone at once.
The common mistake is to treat one as if it were the other. You are told to get out more, join a club, say yes to the invitation. Those can all help with social loneliness. If your hunger is for being known, a fifth evening of making conversation with people who do not know you may leave you more tired and no less alone.
A quick way to tell which one you have
Ask yourself two things. In the last month, have you had somewhere to go, and people to go with? And in the last month, has anyone heard how you actually are, not the “fine, you?” version? If the first is yes and the second is no, what you are missing is depth, not company.
That distinction decides what is worth trying next.
What you can try this week
Neither study was a treatment trial, so none of what follows is a prescription. These are sensible, low-risk ways of interrupting the loop that the findings point towards. Treat them as experiments, and judge them by how they work for you.
Answer one “fine” honestly. Pick one person who has earned it, and replace “fine, you?” with one true sentence: “Honestly, a bit flat this week.” You are not asking them to fix anything. You are giving them something real to respond to. If a face-to-face feels like too much, send it as a message.
Check the threat before you act on it. When the “they meant not now” feeling arrives, ask what you would conclude if a friend described the same message to you. The aim is not to talk yourself out of the feeling. It is to hold it as a hypothesis rather than a fact, and to see whether the evidence agrees.
Choose one deep contact over five shallow ones. A forty-minute walk with one person who knows you is more likely to feed emotional loneliness than a party is. Quality of contact matters more here than quantity, and that fits the whole idea of the research.
Set a small disclosure target. Aim for one thing a week that is slightly more personal than you would normally say, and increase it slowly. The cycle the researchers describe runs on pulling back, so small steps forward matter.
Plan for your own hour. The 20-day study showed loneliness rising and falling through the day. If yours peaks at twenty to ten in the evening, put something there on purpose: a call, a walk round the block, a message you have been saving. An hour you have planned feels very different from an hour that simply arrives.
Tell someone who can help if it feels heavy. If the feeling has lasted for weeks, or sits alongside low mood, poor sleep or thoughts of not wanting to be here, please speak to your GP. In the UK you can also call Samaritans free on 116 123, at any hour.
None of these needs to happen all at once. Pick one. The first attempt may feel awkward, and that is normal.
It is worth saying what a good outcome looks like, because it is rarely dramatic. Nobody will gasp. A friend may simply say “me too”, or “I wondered”, and the evening will feel a little less like glass. One answer like that does not undo months of habit. It does give the loop something it has not met before, which is a reason to expect the next honest sentence to be easier.
If you are waiting for someone to notice
You may have been waiting, perhaps for months, for one of your people to notice that “fine” is not the whole story. That wait is understandable. It is also the move least likely to work, and the research gives a reason why. The loop you are caught in teaches you to expect rejection, and to hide the very evidence that would show the expectation is wrong.
That is not your fault.
It does mean that the way out runs through a small, deliberate act of being known, and that act can start with you. It does not need to be dramatic. It can be one sentence, sent on a Tuesday, to one person.
One truer sentence
You do not have to turn your whole life around this week. The next time someone who matters to you asks how you are, and you feel the word “fine” arriving, try letting a truer one through. Say it in a message if that is easier. Then notice what happens.
Many people are sitting in a lit-up carriage of their own, looking at their own reflection, and wondering whether anyone else has the same view. Some of them are one honest sentence away from finding out that they do. You could be the person who says it first, and the flywheel only knows what you feed it.
Source: Ong, A. D. and colleagues, “Social Asymmetry and Risk of Morbidity and Mortality”, JAMA Network Open (2026), doi 10.1001/jamanetworkopen.2025.58214, a longitudinal study of 7,845 adults over 50 in England; and the companion experience-sampling study of 157 adults in Communications Psychology; both as reported in the Cornell Chronicle, “Why feeling alone may matter more than being alone”, March 2026 (news.cornell.edu/stories/2026/03/why-feeling-alone-may-matter-more-being-alone). The distinction between social and emotional loneliness comes from the wider literature. This article is for information and is not a substitute for medical advice.
Leave a comment