Research · for campaigns, journalists, u3a groups & older people's charities

Common Claims About Loneliness in Later Life, Checked Against the Evidence

Published 2026-11-15

Several things often said about loneliness don’t hold up well against the research. The best-known comparison, with smoking, comes from a 2010 review of social relationships in general, in which only 8 of 148 studies measured loneliness. Living alone is not the same as feeling lonely, older people are not as a group the loneliest, and where the same surveys have been repeated, loneliness among older people has mostly not risen.

These claims are usually made with good intentions, to win attention and funding for something that matters. But they shape how older people are seen and how they see themselves. Someone in their seventies who lives alone, helps run a walking group and rarely feels lonely is told, in effect, that they ought to. And funders who think of loneliness as a problem of old age alone may miss where the evidence points: the very old, people who have recently lost a partner, people in poor health, and younger adults.

8 of 148studies in the 2010 meta-analysis behind the smoking comparison measured loneliness; the rest measured social networks, support, marital status, living arrangements and other aspects of social relationships[1]

Is loneliness as harmful as smoking 15 cigarettes a day?

The comparison traces back to one paper: a 2010 meta-analysis by Julianne Holt-Lunstad, Timothy Smith and Bradley Layton in PLoS Medicine. It pooled 148 cohort studies with 308,849 participants, whose average age at the start was 63.9 years, and asked whether people with stronger social relationships lived longer[1]. Its headline result was an odds ratio of 1.50: across the studies, people with stronger social relationships had 50% greater odds of survival than people with weaker ones. That is a figure for odds of survival; it cannot simply be turned round into a “50% higher risk of death”.

“Social relationships” here meant many things. The paper grouped its measures into functional ones (support people received or felt they could count on, and loneliness) and structural ones (marriage, living alone, the size of social networks, social integration and isolation). Loneliness was measured in only 8 of the studies. The strongest associations came from “complex measures of social integration”, which combine several kinds of social tie and were linked with “a 91% increase in odds of survival”[1].

Where do the cigarettes come in? The authors drew a chart setting their results beside effect sizes from other meta-analyses of risk factors for death. One bar, from a 2008 meta-analysis of smoking, was labelled “Smoking < 15 cigarettes daily”: that is, fewer than 15 a day. Others included alcohol, physical activity and “BMI: Lean vs. obese”. The paper concluded that the effect of social relationships was “comparable with quitting smoking” and exceeded “many well-known risk factors for mortality (e.g., obesity, physical inactivity)”[1]. It did not say that loneliness, or anything else, was equivalent to smoking a set number of cigarettes; it compared the size of statistical associations from different kinds of study.

The figure travelled anyway. “Fewer than 15” became “up to 15”, an association became a risk, and the subject often shifted from social relationships to loneliness. A 2025 paper in Public Health Reports, whose authors include Holt-Lunstad, states that lacking social connection “can increase the risk of premature death comparable to smoking up to 15 cigarettes per day”, citing the US Surgeon General’s 2023 advisory and a 2017 paper by Holt-Lunstad and colleagues[2]. In 2023 she had co-written a commentary in the American Journal of Epidemiology examining the comparison. It noted: “Leaders in health systems, research, government, and popular media have since claimed that the harms of social isolation and loneliness are comparable to that of cigarette smoking.” It concluded that while the analogy had raised awareness, it “often oversimplifies the evidence” and may draw attention to treating individuals rather than to population-level prevention[3].

The underlying research matters: people who are socially isolated or lonely die earlier on average, and our study of loneliness, isolation and earlier death sets out what that evidence can and can’t say, including why the link with loneliness often shrinks once health is taken into account. But “15 cigarettes a day” is a rhetorical comparison, not a finding about loneliness.

Odds of survival by type of social relationship measure, Holt-Lunstad et al. 2010 meta-analysis (higher = stronger link with survival)Weighted average odds ratios for survival from Table 4 of Holt-Lunstad, Smith and Layton (2010), by how social relationships were measured, with the number of studies: complex measures of social integration 1.91 (30 studies); social integration 1.52 (45); all 148 studies combined 1.50; social networks 1.45 (71); loneliness, inverted 1.45 (8); social isolation, inverted 1.40 (8); perceived social support 1.35 (73); marital status 1.33 (62); living alone, inverted 1.19 (17). Not shown: received social support 1.22 (9) and multifaceted measures 1.47 (67).00.511.52Complex measures of socialintegration (30 studies)1.91Social integration (45studies)1.52All 148 studies combined1.50Social networks (71studies)1.45Loneliness (8 studies)1.45Social isolation (8studies)1.40Perceived social support(73 studies)1.35Marital status (62studies)1.33Living alone (17 studies)1.19
Pooled odds ratios for survival (1 = no difference) for people with stronger versus weaker social relationships; loneliness, isolation and living alone are inverted, so higher means a stronger link between not being lonely, isolated or alone and survival. Loneliness was measured in only 8 studies, with scales such as the UCLA Loneliness Scale; its 95% confidence interval was wide (1.08 to 1.94). Two Table 4 categories are not shown: received social support (1.22, 9 studies) and multifaceted measures (1.47, 67 studies). Associations from observational studies, not causes. Source: Holt-Lunstad, Smith and Layton, PLoS Medicine, 2010, Table 4.

Does living alone mean being lonely?

Living alone is a household arrangement; loneliness is a feeling about one’s relationships. Our study of older people living alone shows how widely the share of older people living alone varies between countries, and why that figure on its own tells us little about loneliness.

The research generally finds that people who live alone are more likely to feel lonely than people who live with others, but that most of them don’t. In England, the Office for National Statistics analysed the Community Life Survey for 2016–17, which asked “How often do you feel lonely?”, and found that people living alone were more likely to say “often/always” or “some of the time”[4]. In New Zealand, national home-care assessments of 71,859 people aged 65 and over living at home, with an average age of 82.7, found 29% of those living alone recorded as lonely, against 14% of those living with others[5]. Even in this older group, all being assessed for help at home, about seven in ten people living alone were not recorded as lonely.

In Taipei, a 2019 face-to-face survey of 3,553 older people sorted respondents into groups by loneliness (one question, “do you feel lonely?”, with “sometimes” or “often” counted as lonely), isolation and living arrangement. People who lived alone and were not lonely did no worse on depressive symptoms or life satisfaction than people who were connected and lived with others; the author concluded that “living arrangement is not the determinant to loneliness or isolation”[6].

Context matters too. A cross-national study described in a 2009 review of the myths about older people’s loneliness found that loneliness among people living alone was “higher in places where solitary living was rarest”[7]. And the reverse holds: in the Taipei survey, some of the people who felt lonely lived with others[6].

Share of people aged 65 and over assessed for home care in New Zealand who were recorded as lonely, by living arrangement, 2012-2016National interRAI home care assessments, New Zealand, September 2012 to January 2016, 71,859 people aged 65 and over living at home (average age 82.7): 29% of those living alone were recorded as lonely, against 14% of those living with others. Overall 79% said they were not lonely.0%10%20%30%40%Living alone29%Living with others14%
Loneliness as recorded in a home-care needs assessment (interRAI-HC); the abstract does not give the exact wording. Everyone in the sample was being assessed for help at home, so it is older and in poorer health than older people as a whole. About seven in ten of those living alone were not recorded as lonely. Source: Jamieson et al., Australasian Journal on Ageing, 2018.

Are older people the loneliest age group?

Not as a whole. Our study of loneliness across the life course shows that in several large surveys people in their late 60s and 70s report loneliness less often than young adults, and that loneliness rises again in very old age, following poorer health, the loss of a partner and lower income more closely than age itself.

The pattern is long-standing. Dykstra’s 2009 review in the European Journal of Ageing set out to test the assumption that loneliness is “a problem specifically for older people” and found “only partial support”. Summarising approximately 40 cross-sectional surveys that asked how often people felt lonely, it found loneliness “common only among the very old, that is, those aged 80 and over”, while “the prevalence of loneliness is also high among the youngest respondents, those between the ages of 15 and 24”[7]. In England, the Office for National Statistics found that people aged 16 to 24 reported feeling lonely more often than older age groups, and that people aged 65 to 74 and 75 and over were significantly less likely to be lonely more often than 16 to 24 year olds once other factors were held constant[4].

The BBC Loneliness Experiment, an online survey, gathered answers from 46,054 people aged 16 to 99 in 237 countries, islands and territories, using questions from the UCLA Loneliness Scale. Loneliness “decreased with age”[8]. The authors are clear that people chose to take part, so the sample “is not representative of the broader population”, and that because each person answered once, the pattern may reflect differences between generations rather than change with age[8]. The World Health Organization’s 2025 figures point the same way: it reports that “Between 17–21% of individuals aged 13–29-year-olds reported feeling lonely, with the highest rates among teenagers”[9].

Many surveys sample only people living in their own homes (the Dutch trend analysis in Dykstra’s review, for example, was restricted to people not living in institutions[7]), so they see only part of loneliness among the oldest. And a lower average hides real need: very old age, widowhood and ill health bring loneliness to many people, and those are the groups services most need to find.

Is loneliness among older people rising fast?

The claim sounds plausible: more people live alone, and families are smaller and more spread out. Dykstra’s review lists these reasons, then asks “how much evidence is there for increasing levels of loneliness over time?” Its answer was that the assumption “finds no support. Loneliness levels have decreased, albeit slightly”[7]. The key evidence was a Dutch analysis of over 30 studies of adults aged 18 to 90, spanning 20 years and all using the same 11-item De Jong Gierveld loneliness scale, which found that “loneliness levels have been decreasing over time rather than increasing”. Not every group followed the trend: never-married people aged 65 to 75 showed an increase, while older divorced and widowed people showed decreases[7].

Later and larger studies have broadly agreed. Our study of loneliness trends over time brings together repeated surveys from the United States, Norway, Sweden, Germany and England that asked older people the same question over decades. Most found loneliness among older people holding steady or falling, and later-born generations of older people no lonelier than earlier ones. Where increases have been found among older people, as in pooled analyses of separate studies from China and Europe, the evidence is weaker than one survey repeated over time; other increases have mostly been among younger adults.

As populations age, the number of older people who feel lonely can still grow when the share stays flat. The pandemic years also complicate the picture, because some surveys changed how they were carried out, for example from face-to-face to phone or online interviews, and a jump across such a change is not a trend. But “loneliness is rising fast” is not what the long-running surveys of older people show. Loneliness is an old and common human experience shaped by circumstances, not a new condition of modern old age.

Does the WHO say one in three older people are lonely?

This claim mixes up two different things. In a June 2025 news release, the World Health Organization reported that “1 in 6 people worldwide is affected by loneliness”, a figure for all ages. Its figure for older people is about something else: “While data on social isolation is more limited, it is estimated to affect up to 1 in 3 older adults and 1 in 4 adolescents”[9]. That is social isolation, meaning few social contacts or little social participation, not loneliness, and “up to” signals an upper estimate rather than a typical figure.

The difference matters. A person can see few people and be content, and a person surrounded by family can feel lonely. The same WHO release links loneliness to “more than 871 000 deaths annually”, which is a modelled estimate, not a count; our mortality study explains how it was built.

What the evidence asks of us

None of these corrections weakens the case for action on loneliness; they point it to where help is most needed. For campaigners, services, funders and journalists, the evidence suggests:

  • Retire the smoking comparison, or explain it. The authors of a 2023 commentary, including the lead author of the 2010 meta-analysis, argue that the analogy often oversimplifies the evidence and that more attention should go to “the structures and environments that promote and constrain healthy relationships”[3]. Where the comparison is used, say that it compares effect sizes for social relationships in general, not loneliness.
  • Name the measure. Say whether a figure is loneliness, isolation or living alone, and how it was asked. “Up to 1 in 3 older adults” is isolation; “1 in 6” is loneliness at all ages[9].
  • Target circumstances, not age. Very old age, bereavement, poor health and low income are where loneliness concentrates. Outreach, transport and places to meet aimed at those circumstances may reach more people who feel lonely than services aimed at everyone over 65.
  • Don’t treat living alone as a problem in itself. In the studies here, most older people who lived alone were not lonely, and some people who felt lonely lived with others.
  • Challenge the stereotype. Dykstra’s review noted that even as the Swedish public grew better informed about older people’s circumstances, it kept negative images of old age as a time of boredom, dissatisfaction and loneliness[7]. Groups run by and for older people are among the best evidence against that image.

More studies on loneliness in later life are collected in our loneliness research library.

Language: this piece says "older people" and treats loneliness as a common human experience that anyone can have, not a personal failing or an illness. It keeps apart three things that are often mixed up: loneliness (how people feel about their relationships), social isolation (how few social contacts they have) and living alone (a household arrangement). Where a study uses another term, such as "seniors" or "older adults", we keep it when quoting that study.

Why we're publishing this

The u3a groups, men’s sheds, lunch clubs, senior centres and charities that bring older people together often quote the big claims about loneliness in funding bids and talks. They deserve claims that stand up to a sceptical reader and describe older people as they are: mostly not lonely, often organising things for others, and, where loneliness does strike, affected by circumstances that can be changed. Our overview of the evidence on loneliness in later life summarises the whole library.

How to read this data

The 2010 meta-analysis reports odds ratios for survival from observational cohort studies, pooled across many measures of social relationships: associations, not proof of cause. Loneliness was measured in different ways across the studies cited here: by direct questions such as “How often do you feel lonely?” (England’s Community Life Survey), by questions from the UCLA Loneliness Scale (the BBC Loneliness Experiment), by the 11-item De Jong Gierveld scale (the Dutch trend analysis), and by an item in a home-care assessment (New Zealand). Figures from different measures are not directly comparable. Most age comparisons come from surveys that asked each person once, so they compare age groups, not people growing older. The BBC survey and the New Zealand home-care data are not representative of all older people.

Use this data

Free to cite with attribution to Understanding Aging. u3a groups, men's sheds, senior centres, lunch clubs, befriending services, charities, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Suggested citation: Understanding Aging (2026). Common Claims About Loneliness in Later Life, Checked Against the Evidence. https://understandingaging.org/research/loneliness-myths-evidence

Important: This article summarises published research and statistics for general information. It is not medical or care advice and cannot be used to diagnose, treat or predict anything about an individual. If loneliness, low mood or isolation is affecting your health or someone else’s, a doctor or another health professional can help, and local services for older people can often put you in touch with groups and support nearby.

Sources

  1. Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoS Medicine, 2010. europepmc.org/article/PMC/PMC2910600
  2. Prochnow T, Hawkley LC, Holt-Lunstad J, et al. Clearing the air: using tobacco prevention lessons for social connection. Public Health Reports, 2025. europepmc.org/article/PMC/PMC12583008
  3. Smith RW, Holt-Lunstad J, Kawachi I. Benchmarking social isolation, loneliness, and smoking: challenges and opportunities for public health. American Journal of Epidemiology, 2023. europepmc.org/article/MED/37204190
  4. Office for National Statistics. Loneliness – What characteristics and circumstances are associated with feeling lonely? ONS, 2018. ons.gov.uk/peoplepopulationandcommunity/wellbeing/articles/lonelinesswhatcharacteristicsandcircumstancesareassociatedwithfeelinglonely/2018-04-10
  5. Jamieson HA, Gibson HM, Abey-Nesbit R, Ahuriri-Driscoll A, Keeling S, Schluter PJ. Profile of ethnicity, living arrangements and loneliness amongst older adults in Aotearoa New Zealand: a national cross-sectional study. Australasian Journal on Ageing, 2018. europepmc.org/article/MED/29232761
  6. Hsu HC. Typologies of loneliness, isolation and living alone are associated with psychological well-being among older adults in Taipei: a cross-sectional study. International Journal of Environmental Research and Public Health, 2020. europepmc.org/article/PMC/PMC7764280
  7. Dykstra PA. Older adult loneliness: myths and realities. European Journal of Ageing, 2009. europepmc.org/article/PMC/PMC2693783
  8. Barreto M, Victor C, Hammond C, Eccles A, Richins MT, Qualter P. Loneliness around the world: age, gender, and cultural differences in loneliness. Personality and Individual Differences, 2021. europepmc.org/article/PMC/PMC7768187
  9. World Health Organization. Social connection linked to improved health and reduced risk of early death (news release). WHO, 2025. who.int/news/item/30-06-2025-social-connection-linked-to-improved-heath-and-reduced-risk-of-early-death