What Does 'One in Three Older Adults Are Lonely' Actually Measure

The U.S. Surgeon General’s advisory on social connection, published May 2, 2023, states that approximately half of U.S. adults report experiencing loneliness. That figure will circulate widely over the coming weeks, and it will very likely be applied, informally, to older adults specifically — as shorthand for “half of seniors are lonely.” It should not be. The advisory’s figure describes all U.S. adults, not older adults, and it sits alongside at least two other widely cited numbers for the older population that measure something different again: AARP’s finding that one in three adults 45 and older are lonely, and the National Academies’ finding that roughly one quarter of adults 65 and older are socially isolated. These three figures are not competing estimates of the same thing. They come from different instruments, different age bands, and — critically — two different constructs that get flattened into a single word in most reporting.

Loneliness and isolation are not the same measurement

Isolation describes a structural fact about a person’s network: how many people are in it, how often contact occurs, whether a person lives alone. Loneliness describes a subjective state: whether a person feels the gap between the connection they have and the connection they want. A person can be surrounded by contact and still feel lonely; a person can have a small network and not feel lonely at all. The two are correlated but not interchangeable, and the National Academies’ 2020 consensus report is explicit that they require separate assessment tools rather than one omnibus question.

This distinction is not academic pedantry. It changes what a headline number can support. AARP’s 2018 survey of 3,020 adults aged 45 and older used the 20-item UCLA Loneliness Scale — a validated instrument with a long history in the psychological literature, which is why AARP’s “one in three” figure is directly comparable to academic work using the same scale. It measured loneliness: a felt state. The result was that 33% of respondents who reported speaking with neighbors were lonely, against 61% of those who had never spoken with a neighbor — a striking gradient, but one describing subjective experience, elicited through a structured battery of statements like “I feel left out” and “I feel isolated from others,” not a count of social contacts.

The National Academies’ “roughly one quarter” figure describes something structurally different: social isolation, assessed through indicators of network size, contact frequency, and living arrangement, among adults 65 and older specifically — a narrower and older cohort than AARP’s 45-plus sample. The 2020 clinical commentary on the report, published in the American Journal of Geriatric Psychiatry, underscores that this is precisely why the National Academies pushed for isolation to be assessed as a distinct clinical variable, not folded into a loneliness screen, if health systems are going to act on it.

So: one in three (45+, loneliness, UCLA scale) and one in four (65+, isolation, structural indicators) are both true, and neither one is a more precise version of the other. They answer different questions about different, overlapping populations.

Where the Surgeon General’s “half” figure fits

The advisory’s headline that about half of U.S. adults report loneliness is a whole-population figure, and the advisory itself frames the burden as heaviest and best documented in specific groups — including older adults living alone — rather than presenting loneliness as uniformly distributed by age. Applying the population-wide half figure to older adults specifically, without checking which slice of “adults” produced it and under what instrument, repeats the same error that makes the AARP and National Academies numbers look contradictory when they are not. The advisory is a synthesis document assembling many prior studies, several of which appear elsewhere in this review’s source list, and its top-line number should be read as a description of the adult population as a whole, not as a substitute for the age-specific figures already available.

The AHA statement adds a third axis: perceived versus objective

The American Heart Association’s 2022 scientific statement on social isolation and cardiovascular and brain health complicates the picture further by distinguishing objective isolation — the structural absence of contact — from perceived isolation, which is closer to loneliness but assessed through a different set of instruments than the UCLA scale. The statement’s headline association, a roughly 30% increased risk of heart attack, stroke, or death from either among isolated or lonely individuals, draws on studies using both objective and perceived measures without fully separating their contributions. The statement itself names this as a gap: it explicitly identifies the absence of intervention evidence, and by extension the difficulty of comparing across measurement approaches, as the central problem facing the field. A newsroom summary of the same statement notes that older adults face elevated risk of isolation specifically, but the underlying instruments used to establish that risk vary from study to study within the meta-analytic base, which is standard practice in cardiovascular epidemiology but makes cross-study prevalence comparison unreliable.

Why the instrument, not just the number, matters

Holt-Lunstad’s 2015 meta-analysis in Perspectives on Psychological Science is useful here because it reports isolation, loneliness, and living alone as three separate risk factors with three separate odds ratios — 1.29, 1.26, and 1.32 respectively for early mortality — rather than collapsing them. That paper’s authors evidently thought the distinction was worth preserving even though the three factors move together in most populations. If a meta-analysis built specifically to establish mortality risk keeps these constructs apart, a policy brief or news article that merges “lonely,” “isolated,” and “alone” into a single interchangeable statistic is doing less careful work than the primary literature it is drawing from.

The practical consequence for anyone using these figures: a program built to increase contact frequency — a friendly visiting scheme, a transportation service to get isolated seniors to a community center — is targeting isolation as the National Academies defines it. It may or may not move the loneliness figure AARP measured, because loneliness is not simply the inverse of contact frequency. Conversely, an intervention aimed at loneliness — a therapeutic or peer-support program addressing how a person interprets and feels about their social world — may do little to change someone’s actual isolation as a structural fact. Programs justified by citing “one in three older adults are lonely” should be evaluated against loneliness outcomes, not contact counts, and vice versa. Conflating the two invites evaluating a program against the wrong number entirely.

What would resolve the ambiguity

A study following the same older adults over time with both a validated loneliness scale and a structural isolation index, administered together rather than in separate surveys years apart, would let researchers see how much the two measures actually diverge within the same people — rather than inferring the difference by comparing separate samples, separate years, and separate age cutoffs, as the current literature requires. Until that exists, any single headline percentage about older adults and connection should prompt one question before it is repeated: which of the two things was actually measured, and in whom.