UCLA Loneliness Scale (Version 3): Take It, Score It, Understand Your Results
The UCLA Loneliness Scale is a self-report measure of perceived loneliness developed at UCLA in 1978 and refined into its current Version 3 by Daniel Russell in 1996. Version 3 has 20 items rated on a four-point scale (never / rarely / sometimes / often), with a total range of 20 to 80, and roughly half the items are positively worded and reverse-coded so that higher scores always mean more loneliness. It is one of the most widely cited loneliness measures in social-psychology and population-health research worldwide, used by the CDC, by the US Surgeon General in the May 2023 advisory on loneliness, and in the WHO's November 2023 framing of loneliness as a global public health concern. Critically, the UCLA Loneliness Scale is a research and wellness measure — not a clinical screen and not a diagnostic instrument. Russell's 1996 paper does not publish clinical cutoffs, and a high score on its own is not a diagnosis.
What is the UCLA Loneliness Scale?
The UCLA Loneliness Scale is a self-report measure of perceived loneliness developed at the University of California, Los Angeles. The original 20-item scale was created in 1978 by M.L. Ferguson, Daniel Russell, and Letitia Anne Peplau. It was revised in 1980 as the R-UCLA, and Russell published the current canonical form, Version 3, in 1996 in the Journal of Personality Assessment. A 3-item short form was developed by Hughes, Waite, Hawkley, and Cacioppo in 2004 for use in the Health and Retirement Study and other large population surveys. Across its versions, the UCLA scale has become the most widely cited unidimensional loneliness measure in social-psychology, gerontology, and population-health research worldwide. The scale is framed by its author as a measure of perceived loneliness, not as a clinical instrument.
What the UCLA Loneliness Scale measures
The UCLA Loneliness Scale measures the subjective experience of loneliness — feelings of disconnection, isolation, lack of companionship, and rejection — through 20 items that capture how often a respondent feels these states. Items are framed both positively (for example, asking about feeling part of a group of friends) and negatively (for example, asking about feeling left out), and the balanced wording requires reverse-coding the positive items so that higher scores always reflect more loneliness. Russell's 1996 validation reported Cronbach's alpha between .89 and .94 across four samples — college students, nurses, teachers, and elderly adults — and a one-year test-retest correlation of .73, indicating strong reliability. The scale measures perceived loneliness rather than objective social isolation; the two constructs are distinct and require different instruments.
How the UCLA Loneliness Scale is administered
Version 3 is a self-report questionnaire that takes approximately three to five minutes to complete and is written at a sixth- to eighth-grade reading level, making it accessible to a broad range of respondents. It has 20 items, each rated on a four-point Likert scale (1 = never, 2 = rarely, 3 = sometimes, 4 = often). Roughly half the items are phrased positively and must be reverse-coded before scoring. After reverse-coding, scoring can be expressed two valid ways: summing the items for a total in the 20 to 80 range, or averaging the items for a mean in the 1 to 4 range. Russell 1996 does not publish a clinical cutpoint — higher scores simply indicate greater perceived loneliness. The 3-item short form (Hughes 2004) uses a 1 to 3 response scale and totals 3 to 9, useful for brief surveys. No special preparation is needed.
Who uses the UCLA Loneliness Scale
The UCLA Loneliness Scale is used in social-psychology and gerontology research worldwide, in population-health surveys, and in clinical trials of community-engagement and therapy interventions. The CDC's social-connectedness work and AARP's older-adult loneliness programming both reference the broader loneliness measurement landscape that the UCLA scale anchors. The 3-item Hughes 2004 short form is the version embedded in the US Health and Retirement Study and in the English Longitudinal Study of Ageing, and is favored by surveys that need brief loneliness screening. The US Surgeon General's May 2023 advisory on loneliness and the World Health Organization's November 2023 declaration of loneliness as a global public health concern both rely on the broader research base that uses the UCLA scale. It is less common in clinical psychiatry because there is no diagnostic use case, but it is increasingly used by researchers studying social determinants of health.
UCLA Loneliness Scale is a research instrument, not a diagnosis
The UCLA Loneliness Scale measures perceived loneliness — it does not diagnose any clinical condition, and Russell's 1996 Version 3 paper does not publish clinical cutoffs. Any numerical bands you may encounter online are third-party conventions, not validated thresholds. Loneliness itself is not a clinical disorder, although CDC notes that loneliness and social isolation are associated with depression, anxiety, suicidality, cardiovascular disease, and earlier death. A high UCLA score should be read as a personal signal — a prompt to reflect on social connection rather than evidence of a medical problem. If your loneliness is persistent and is paired with low mood, sleep changes, or difficulty functioning, a conversation with a primary care clinician or mental health professional is appropriate; therapy is effective in roughly ten to twenty weeks for loneliness with addressable underlying factors. If you are in crisis, the 988 Suicide & Crisis Lifeline is a universal US resource you do not need a screening result to call.
How to interpret your UCLA Loneliness Scale score
The UCLA Loneliness Scale is a measure of felt loneliness — the felt gap between the social contact a person has and the social contact they want. It is a research and wellness tool, not a clinical screen, and there is no published clinical cutoff that maps a score to a diagnosis. Higher scores mean more felt loneliness; lower scores mean less.
The most-used modern form is Version 3, published by Daniel Russell in 1996. It has 20 items rated on a four-point scale (1 = never, 2 = rarely, 3 = sometimes, 4 = often). Roughly half the items are phrased positively, and those must be reverse-coded before scoring so that higher numbers always mean more loneliness.
Two valid ways to express the same score
Once items are reverse-coded, scoring goes one of two ways:
- Sum the 20 items for a total in the range 20–80, where 20 means lowest loneliness and 80 means highest.
- Average the 20 items for a mean score in the 1–4 range — this is the format SPARQ Tools at Stanford uses.
Both are valid presentations of the same response pattern. If you see one form online and a different form in a paper, divide or multiply by 20 to translate.
Why no canonical “moderate / high / severe” bands exist
Russell’s 1996 Version 3 paper backs the scale’s reliability and factor structure, but does not publish clinical cutpoints. The bands that circulate online (a “28-43 moderate / 44-55 moderately high / 56+ high” pattern, for example) are third-party conventions, not Russell’s, and have no canonical source. The honest read is a comparative one: your score next to a prior result, next to a peer group in a study, or next to your own sense of how connected you feel.
The 3-item short form (Hughes et al. 2004) uses different scoring: each item is scored 1 = hardly ever, 2 = some of the time, 3 = often, summed for a total range of 3–9. Large population surveys often split at ≥6 as “lonely,” but Hughes 2004 itself did not publish a clinical cutoff — that split is a research convention, not a clinical threshold.
Version 3 vs the 3-item short form: which one are you taking?
Many searches that land on “UCLA Loneliness Scale” pages are about one of two different instruments. They share a lineage but differ in length, response scale, total range, and intended use.
| Feature | Version 3 (Russell 1996) | 3-item short form (Hughes 2004) |
|---|---|---|
| Number of items | 20 | 3 |
| Response scale | 4-point (never / rarely / sometimes / often) | 3-point (hardly ever / some of the time / often) |
| Total range | 20–80 (or 1–4 averaged) | 3–9 |
| Time to complete | About 3–5 minutes | Under a minute |
| Internal consistency (Cronbach α) | .89–.94 across four samples | .72 |
| Best used for | Detailed individual measurement, research with adequate time, repeat measurement | Telephone interviews, large population surveys, brief screening in primary care |
The 3-item form draws its items directly from the 20-item Revised UCLA Loneliness Scale: how often you feel you lack companionship, how often you feel left out, and how often you feel isolated from others. In Hughes and colleagues’ development paper, the 3-item form correlated r = .82 with the full 20-item version (n = 229). That correlation is high enough to make the short form a fair stand-in when interview time is short, but Version 3 catches more nuance and is the standard when time allows.
A practical implication: if the “UCLA Loneliness Scale” you took online had three questions, it was the Hughes 2004 short form, with scores in the 3–9 range. Twenty questions means Version 3, with scores in the 20–80 range (or 1–4 if averaged).
What the UCLA scale measures — and what it doesn’t
The UCLA Loneliness Scale measures felt loneliness: the inner sense of feeling cut off, left out, or short on close company. It does not measure social isolation, which is a different idea.
The CDC tells the two apart clearly. Social isolation is structural — “a person does not have relationships or contact with others and has little to no social support.” Loneliness is the inner sense of “feeling alone or disconnected from others,” and the two can show up on their own. A person can be socially active and still feel lonely; another can live alone and not feel lonely. Hughes and colleagues made the same point in their 2004 paper, noting that “objective and subjective measures of social isolation tap different aspects of social experience”. The UCLA scale sits firmly on the inner-sense side.
Unidimensional, by design
Russell’s 1996 factor analysis backed treating Version 3 as a one-construct measure of loneliness, with two method factors that adjust for whether items are positively or negatively worded. In practice that means a single overall score is the intended output — the scale is not built to give you separate sub-scores for “social loneliness” versus “emotional loneliness.”
Some researchers see that as a limit. The UCLA scale “has been criticised for being unidimensional,” because some frameworks split social loneliness from emotional loneliness; researchers wanting that split typically use De Jong Gierveld or SELSA instead. The Campaign to End Loneliness frames the UCLA take in lay terms as the gap between “the social contact we have, and the social contact we want” — one number, one core idea.
What the scale does not do
- It does not diagnose any clinical condition. There is no DSM or ICD diagnosis of “loneliness.”
- It does not split social from emotional loneliness — for that, a multi-construct scale is the fit.
- It does not measure how many people you know or how large your social network is — those are isolation measures, not loneliness measures.
- It does not predict any one person’s future health from a single take; it captures one moment in time.
How the UCLA Loneliness Scale compares to other loneliness measures
The UCLA scale is the default in loneliness research — a 2001 metastudy of 149 studies found it used in about 27%, “far more than for any other formal scale,” and it stayed the most-used one-construct loneliness scale worldwide as of 2018. But it is not the only option, and each option has a different best use.
| Scale | Items | Construct | Best used when |
|---|---|---|---|
| UCLA Loneliness Scale Version 3 | 20 | One-construct felt loneliness | You want a full per-person measure with strong psychometrics |
| UCLA 3-item (Hughes 2004) | 3 | One-construct felt loneliness | Large surveys, phone interviews, brief screening |
| De Jong Gierveld | 6 | Multi-construct — splits social vs emotional | You need to know if someone lacks a broad network or a close bond |
| ONS direct question | 1 | Direct self-label | You have one slot in a survey and want a single direct measure |
The UK’s Office for National Statistics recommends pairing both direct and indirect measures where possible, because “the stigma of loneliness may mean that people underreport their feelings if asked directly”. Indirect items like the UCLA 3 can pick up loneliness that a single direct question misses. For a person comparing their own loneliness across time, the choice is simpler: pick one scale, use it consistently, and watch the trajectory rather than any single number.
When to talk to a clinician about loneliness
The UCLA Loneliness Scale is not a clinical screen, and a high score is not a diagnosis. That said, persistent loneliness can co-occur with depression, anxiety, and other conditions that benefit from professional care, and CDC lists loneliness and social isolation as risk factors associated with depression, anxiety, suicidality, cardiovascular disease, and earlier death.
When loneliness has roots you can act on, therapy “typically spanning ten to twenty weeks” tends to be effective, with cognitive-style work that targets thought patterns showing the strongest results. A talk with a primary care clinician or mental health pro can sort out whether your loneliness is part of a wider picture worth care or a tight experience tied to one life event.
Consider scheduling a conversation with a clinician if any of these apply:
- Your loneliness has lasted weeks or months and is not shifting, especially if paired with low mood, hopelessness, or loss of interest in things you used to enjoy
- Sleep, appetite, energy, or concentration have changed alongside the loneliness — these are signals that depression or anxiety may be involved
- You have lost a major relationship recently — bereavement, divorce, a move, or an empty-nest transition — and the loneliness is not easing as weeks pass
- You are in a higher-risk group named by CDC, such as living alone, being a recent immigrant, being a young or older adult, or facing limited resources
- You are an older adult — community programs like those AARP offers can add to, not replace, a primary care visit when loneliness sticks around
- A different mental-health screen flags something — for example, if a brief depression screen like the PHQ-9 suggests depressive symptoms, that is a stronger reason to make an appointment than a UCLA score alone
A clinician can help tell loneliness as a feeling apart from loneliness as one part of a treatable condition, and link you to resources — therapy, peer support, community programs — that the scale itself cannot.
If you are in crisis
If you are having thoughts of suicide or self-harm, or are in immediate danger, call or text the 988 Suicide & Crisis Lifeline in the United States. In a life-threatening situation, call 911. 988 is a universal resource and you do not need a positive screen on any instrument to use it.
Frequently asked questions
What is the UCLA Loneliness Scale?
The UCLA Loneliness Scale is a self-report measure of felt loneliness built at UCLA. The current standard is Version 3, published by Daniel Russell in 1996, with 20 items rated on a four-point scale. It is a research and wellness tool, not a clinical screen, and does not diagnose any condition.
What is the difference between Version 3 and the 3-item short form?
Version 3 has 20 items rated 1–4, with a total range of 20–80. The Hughes 2004 short form has 3 items rated 1–3 (hardly ever, some of the time, often), with a total range of 3–9. The short form is intended for large surveys and telephone interviews; Version 3 is the standard when time is available.
Where can I find the official PDF?
The UCLA Loneliness Scale Version 3 is hosted by Stanford’s SPARQ Tools, and the 3-item Hughes 2004 form appears in dozens of open-access papers and population-survey codebooks. We do not host an interactive UCLA scale on this page.
How long does it take to complete?
Version 3 takes about 3 to 5 minutes and has been validated across general adult populations and several specific subgroups. The 3-item short form takes well under a minute.
Is this a diagnostic test?
No. The UCLA Loneliness Scale measures felt loneliness as a research and wellness tool, and Russell’s 1996 paper does not publish clinical cutoffs. Loneliness itself is not a clinical diagnosis, though it often shows up with conditions like depression and anxiety that a clinician can review on their own.
Who developed it and when?
The original 20-item UCLA Loneliness Scale was developed by M.L. Ferguson, Daniel Russell, and Letitia Anne Peplau at UCLA in 1978. It was revised in 1980 (R-UCLA) and again in 1996 (Version 3, by Russell). The 3-item short form was developed by Hughes, Waite, Hawkley, and Cacioppo in 2004 for the Health and Retirement Study.
How does it compare to the De Jong Gierveld scale?
The UCLA scale is one-construct — it gives one overall loneliness score. The De Jong Gierveld scale splits social loneliness (lacking a broad network) from emotional loneliness (lacking a close intimate bond). The right pick depends on whether you want one number or two.