The UCLA Loneliness Scale Was Built Before the Internet, and It Shows

When the AARP Foundation reported in 2018 that one in three U.S. adults aged 45 and older were lonely, the figure came from a 20-item instrument that has been in continuous use since the late 1970s: the UCLA Loneliness Scale. When Cigna reported in 2020 that 61% of U.S. adults were lonely, the figure came from a different, shorter set of questions built for that survey. Both numbers get quoted as if they measure the same thing. They do not, and the gap between them is a good occasion to ask what the older instrument actually assumes about social contact, and whether those assumptions still hold.

What the scale asks

The UCLA Loneliness Scale asks respondents to rate, on a four-point scale, how often they feel things like “I have nobody to talk to,” “I feel left out,” or “I feel part of a group of friends.” It does not ask how many people someone talked to yesterday, or how many friends live within driving distance. It is explicitly a measure of the subjective experience of loneliness, not of the structural fact of isolation — a distinction the field has spent decades trying to keep straight and that popular coverage routinely collapses.

That design choice reflects the theoretical position, later developed at length by John Cacioppo and William Patrick, that loneliness functions as an internal signal rather than a description of one’s social calendar. Cacioppo’s 2008 account frames loneliness as an aversive state analogous to hunger, evolved to motivate reconnection regardless of how many people are objectively nearby. Two people with identical numbers of daily interactions can score very differently on the scale, and the scale is built to make room for exactly that. This is also why the National Academies’ 2020 consensus report on isolation in older adults treats loneliness and isolation as separate, independently predictive constructs rather than two measures of one thing, and why Holt-Lunstad’s 2015 meta-analysis reports separate mortality odds ratios for each — 1.26 for loneliness, 1.29 for social isolation, 1.32 for living alone.

The instrument’s virtue is comparability. Because AARP used the same 20-item scale that has appeared in the academic literature since the 1970s, its 2018 finding sits inside a long run of comparable studies rather than standing alone. That is a genuine methodological strength, and one reason the AARP figure carries more weight in the research literature than headline survey numbers built from a single bespoke question.

What it assumes about contact

The scale’s items were written before telephone calls were cheap, before email, and decades before video calling or messaging apps existed as a form of daily contact. Its questions about feeling “part of a group of friends” or having “people I can turn to” do not specify a channel. Respondents supply their own sense of what counts, and that is precisely where the instrument’s age becomes a live methodological question rather than a historical footnote.

A respondent in 1978 answering “I have nobody to talk to” was almost certainly evaluating in-person or telephone relationships, because there was little else. A respondent in 2021 answering the same item might be counting a group chat that runs constantly, a parasocial relationship with an online community, or a daily video call with a parent in another state — all of which are real forms of contact, none of which the scale’s authors could have anticipated, and none of which the instrument distinguishes from in-person contact in the response options. The scale asks about the feeling, not the medium, which was defensible when the media available were limited and roughly equivalent in intimacy. It is a different proposition now, when the available media vary enormously in how much they resemble face-to-face contact.

This matters because the Harvard Making Caring Common survey from February 2021 found 61% of young adults aged 18 to 25 reporting serious loneliness, and about half of lonely young adults saying no one had spent more than a few minutes recently asking how they were doing in a way that felt genuine. That last finding is instructive: it suggests the deficit young adults are reporting is about the depth or genuineness of contact, not its frequency or channel. A generation that communicates constantly by text can still describe an absence of the kind of exchange the UCLA scale is trying to capture. The instrument was arguably built for exactly this distinction — feeling versus frequency — but it was not built anticipating that frequency itself would become nearly unlimited while feeling stayed scarce. Whether high-frequency, low-depth digital contact moves someone’s score on the scale in either direction is an empirical question the instrument was never designed to answer, because the scenario did not exist when it was written.

The instrument-mismatch problem in practice

This is why the AARP and Cigna figures cannot simply be averaged or reconciled. AARP’s 33% used a validated multi-item scale with decades of prior use, applied to adults 45 and older. Cigna’s 61% used a different set of items, applied to all U.S. adults, and showed a working-age population — 73% of workers aged 18 to 22 — reporting loneliness at more than twice the AARP rate for older adults. Some of that gap is real: younger adults report more loneliness across nearly every survey that has asked, including the Harvard study. But some of the gap is instrument, not population. Comparing a 20-item validated scale to a survey house’s proprietary items is comparing two different measuring devices and reporting the disagreement as a finding about people.

Holt-Lunstad’s 2021 review, arguing for social connection as a modifiable prevention target alongside diet and smoking, depends on being able to track a single construct over time and across studies. That kind of prevention framework needs an instrument stable enough that a change in the score reflects a change in the person, not a change in the technology surrounding them. The UCLA scale’s stability over 40-plus years is what makes it useful for that purpose. It is also what makes its silence on channel increasingly consequential the longer it is used without revision.

What would resolve this

None of this means the scale should be discarded; its track record is exactly what gives the AARP figures their credibility. But a stronger instrument, or a stronger companion measure, would ask separately about the channel and intimacy of a person’s contacts — in-person, voice, video, text — rather than leaving respondents to average across all of them when they answer a single item like “I feel part of a group of friends.” That would let researchers test directly whether digitally mediated contact substitutes for in-person contact in reducing loneliness scores, complements it, or does neither. At present, the literature can report that loneliness is high and that digital communication is widespread, but it cannot yet report, from this instrument, how much of one is doing anything to the other. That is not a small gap. It is close to the central question anyone using this scale to study a digitally mediated population actually wants answered.