How loneliness is measured, and what a loneliness score can say

What loneliness research measures

Start with the thing being measured, because it is narrower than the word. The World Health Organization describes loneliness as the painful feeling that arises from a gap between desired and actual social connections, while social isolation refers to the objective lack of sufficient social connections. Those are two different measurements. One asks how your relationships feel against what you wanted from them. The other counts.

That is why a loneliness questionnaire never asks how many friends you have. Someone with a full diary can report a wide gap; someone who sees three people a month can report none. The gap is the subject, and the gap is subjective. That is the point of the measure and also the reason it is uncomfortable to take.

Public health bodies treat the feeling as a health matter rather than a mood. The report of the WHO Commission on Social Connection, published in 2025, finds that 1 in 6 people worldwide is affected by loneliness, and links loneliness to an estimated 100 deaths every hour, more than 871 000 deaths a year. Those are statements about populations. They are not a statement about you, and nothing here turns a population figure into a fact about one person.

The UCLA Loneliness Scale, and what it actually is

Most loneliness research is built on the UCLA Loneliness Scale. The current edition, usually written as Version 3, is described in the SPARQtools measure library as a 20-item measure that assesses how often a person feels disconnected from others, with each statement answered on a four-point rating scale running from never to always.

Two features of it matter for anyone meeting a copy of it online. The first is reverse wording. Some of the 20 statements are put the positive way round, and the same entry says researchers later reverse-code the positively worded items so that high values mean more loneliness, and then calculate a score for each respondent by averaging their ratings. That is not a trick played on the reader. It is there so that answering every statement the same way does not, by itself, produce a result.

The second is what the score is not. The published description sets out the items, the response scale and the scoring, and carries no clinical cut-off and no diagnostic threshold. So if you meet a copy of this scale that prints a total out of 80 with a severity label under it, the label came from whoever built that page. It is not part of the instrument being copied, and it is the single most common thing wrong with a free loneliness test.

The three-item short form, and the four questions a national survey asks

A 20-item questionnaire is too long to bolt on to a national survey that is mostly about housing or employment, so a three-item version is what official statistics actually use. In the United Kingdom the Office for National Statistics recommends four questions to capture different aspects of loneliness, and says the first three questions are from the University of California, Los Angeles (UCLA) three-item loneliness scale.

The four, with the answers offered:

• How often do you feel that you lack companionship? Hardly ever or never, some of the time, or often.
• How often do you feel left out? The same three answers.
• How often do you feel isolated from others? The same three answers.
• How often do you feel lonely? Often or always, some of the time, occasionally, hardly ever, or never.

The Government Analysis Function harmonised standard sets out the arithmetic for the first three: hardly ever or never equates to one, some of the time to two, and often to three, and the three are then summed to create a total for each individual. The ONS guidance gives the range that produces, stating that the lowest possible combined score on the loneliness scale is 3 and the highest is 9.

That is the whole instrument. Three questions, three answers each, one total between 3 and 9, plus a fourth question that is kept separate.

Why the direct question is asked as well, and why it under-reports

On the face of it the first three questions make the fourth unnecessary. The ONS explains the split: some measures ask about loneliness directly while others ask about emotions associated with loneliness from which loneliness is then inferred. The first three are the indirect kind, and the word lonely does not appear in any of them.

The Government Analysis Function says why that is useful. The first three questions in the harmonised standard measure loneliness indirectly, without mentioning the term, and that measure is beneficial when there is potential for non-declaration of loneliness, a phenomenon thought to be more prevalent in males than females. The ONS makes the same point in plainer words: there is variation in how people understand the term loneliness, and some people might be reluctant to admit to loneliness, and this might be particularly true of certain groups such as older men.

The Campaign to End Loneliness states the reason for keeping both kinds together: the stigma of loneliness may mean that people underreport their feelings if asked directly.

So the direct question is known to be answered conservatively, and it is asked anyway. The reason is that it is the only one of the four where the person decides for themselves whether the word applies to them. The indirect three infer a feeling from answers about companionship and being left out; only the fourth lets someone say it. A measurement programme that dropped the direct question would lose that, and one that kept only the direct question would miss the people the word puts off. Keeping both is a decision about honesty rather than a redundancy, and it is worth knowing that the people who run the national measure say out loud that their own question under-counts.

What a total between 3 and 9 is actually for

It is for comparing groups of people, and it is good at that. Ask the same three questions of thousands of people in the same wording every year and you can see whether one year differs from the last, whether one region differs from another, or whether an age group has moved. That is what a national indicator is: a number designed to be compared with other instances of itself.

What three questions cannot do is tell one person much. Each has three answers, so the instrument has seven possible totals in total. Moving from some of the time to often on one question shifts the total by exactly as much as moving on a different question about a different thing. The resolution is right for a survey of thousands and wrong as a reading of you.

And none of the published guidance read for this page attaches a clinical threshold to an individual's total. Loneliness is not a diagnosis. There is no illness for a questionnaire to find, no cut-off that decides anything, and nothing a score can hand you that a clinician would act on by itself. What these scales record is a feeling, reported by the person having it, on the day they were asked.

Why the self-check here reports four patterns and no total

The loneliness self-check on this site is not the UCLA scale and contains none of its items. Its sixteen statements are original, written from the public definitions of loneliness published by the WHO Commission on Social Connection and the Campaign to End Loneliness. That constraint is the point rather than a shortcut: copying a research instrument into a browser tab and printing a band underneath it is the practice described two sections above.

The four patterns it reports separately are these.

• Unseen in company. Being among people without being known by them. This one usually turns on the kind of conversation a room allows, not on how many people are in it.
• Missing someone to confide in. Having people to do things with, and nobody to hand a hard thing to.
• The gap between contact wanted and had. The pattern the research definitions are really about, and the one that moves from either side: some weeks the contact drops, some weeks what you want from it rises.
• Withdrawing when it hurts. Turning down the invitation at the point you would most benefit from it. The only one of the four you can test differently on the same evening you notice it.

Adding those four together would produce one number that hides which of them moved. A full diary with nobody to confide in and an empty diary with one close friend in it are different weeks needing different things, and they can average to the same figure. So there is no total, no score out of anything, no severity band, and no comparison with other people. The four meters are the result, and the note under each describes only how that pattern showed up in your own answers today.

Sixteen questions answered on one day also cannot count the people in your life, judge your relationships, or say anything about your health. Loneliness moves with a move to a new city, a shift pattern, an illness, a bereavement, a break-up, a new baby. A hard month reads differently from a settled one, and that is a fact about the month.

If this is heavy to read

Some of those questions sit close to the bone, and reading about measurement does not make the thing being measured any lighter. Two plain things are worth saying before you go back to it.

If you are thinking about harming yourself, or you do not feel safe, stop here. Contact emergency services or a crisis line now. In the US and Canada, call or text 988. In the UK and Ireland, call Samaritans free on 116 123. Anywhere else, findahelpline.com lists verified lines by country.

And if a low mood, or a sense of being apart from people, has stayed with you for weeks, that is worth saying out loud to a doctor or to someone you trust. Loneliness and low mood arrive together often enough that telling them apart is a job for a person rather than a questionnaire. Nothing on this site is a diagnosis, and nothing on it is a reason to wait.