Technology & Social MediaMethods & Data
The Instrument That Cannot See a Screen
National loneliness measurement rests on instruments that were designed to ignore how contact happens. That design choice is defensible for validity and disabling for technology policy.
Center for Social Connection

The most durable element of the UK’s 2018 strategy A Connected Society was not a programme. It was a measurement mandate: loneliness questions were embedded into Office for National Statistics collection, alongside funding for social prescribing. That made the UK the first government to commit to a recurring national loneliness series, and it forced a choice of instrument. The battery that emerged is built around a direct question about how often a person feels lonely, sitting next to items adapted from the UCLA Loneliness Scale.
Those UCLA items carry an assumption that almost never gets stated in policy documents. They do not ask how contact happens.
Channel-agnostic by design
The UCLA scale asks whether a person lacks companionship, whether there are people they can turn to, whether they feel understood. The 20-item version predates broadband, mobile phones, and social platforms, and nothing in its wording had to be updated for any of them, because no item ever specified a medium. A respondent who feels understood by three friends they text daily and a respondent who feels understood by three neighbours they see in person score identically.
This is defensible, and the theory behind it is coherent. John Cacioppo and William Patrick’s 2008 account frames loneliness as an aversive signal — analogous to hunger — that registers a perceived deficit between the relationships a person has and the ones they want, and that tracks measurable changes in stress physiology, sleep, and immune function. A signal of that kind responds to the felt sufficiency of relationships, not to their delivery mechanism. An instrument that scored a video call differently from a coffee would be measuring something other than the signal.
The consequence for technology policy is severe. A channel-agnostic instrument cannot, in principle, answer whether mediated contact substitutes adequately for in-person contact. It can only report the resulting subjective state. When national loneliness figures are used to argue about smartphones, they are being asked a question their instrument was constructed not to ask.
That is the central weakness in the empirical fight over adolescent technology use. Jonathan Haidt’s 2024 argument that a shift from play-based to phone-based childhood drove rising anxiety, depression, and loneliness from the early 2010s onward is a claim about channel substitution. The collaborative review documents Haidt maintains with Jean Twenge compile studies on both sides, and the dissent is real. But the loneliness half of the evidence base is largely built from instruments that record only the outcome state, so it can establish that reported loneliness rose while phone use rose, and cannot separate that from the several other things that changed in the same decade.
The enumerating instruments count something else
The structural family of measures does the opposite. Social isolation indices count household composition, marital status, group memberships, and contact frequency. Robert Putnam’s Bowling Alone took this approach to a national scale in 2000, tabulating club membership and informal socialising over decades and distinguishing bonding from bridging social capital. The National Academies’ 2020 consensus report used structural criteria to estimate that roughly one quarter of adults aged 65 and older are socially isolated.
These measures are not interchangeable with loneliness measures, and the 2015 meta-analysis by Julianne Holt-Lunstad and colleagues in Perspectives on Psychological Science is the clearest demonstration. Social isolation carried an odds ratio of 1.29 for early mortality, loneliness 1.26, and living alone 1.32 — three separate coefficients for three separate constructs, all surviving adjustment for health status, and all more predictive in samples averaging under 65. Structural and subjective measures are independently predictive. They are also constantly conflated in secondary coverage.
Enumerating instruments could in principle be made channel-sensitive, because they count events. In practice most were written when contact meant physical contact, and they inherit that. A weekly visit and a weekly hour on a group video call are not scored the same way by most network indices, and no widely used index has settled the question of whether they should be.
Same population, three numbers
The instrument problem is easiest to see where the population is held roughly constant and only the question changes.
Cigna’s 2020 workplace report put the share of U.S. adults sometimes or always feeling lonely at 61%, up seven points year over year. Harvard’s Making Caring Common survey in February 2021 reported 36% with serious loneliness, rising to 61% among 18-to-25-year-olds. The Surgeon General’s 2023 advisory summarised the position as approximately half of U.S. adults reporting loneliness, and set the mortality comparison at smoking up to 15 cigarettes a day. Gallup’s 142-country work with Meta, fielded June 2022 to February 2023, found 24% very or fairly lonely and 49% not lonely at all, while separately reporting that 72% felt very or fairly socially connected. Gallup’s own follow-up analysis makes the point that connectedness and loneliness are not simple inverses. Its October 2024 update found more than one in five adults worldwide lonely a lot of the previous day.
None of these figures is wrong. They are answers to different questions: a two-point frequency threshold, a severity threshold, a global self-label, a yesterday-experience item. The 2023 BMC Public Health review of the field named inconsistent measurement as the principal barrier to comparing findings across studies, and the spread from 21% to 61% is what that barrier looks like in numbers.
Date discipline matters here too. The CDC’s federal surveillance estimates appeared in MMWR in June 2024 but describe 2022 data, and the advisory’s synthesis draws on collection that mostly predates it.
The remote work case
Gallup’s 2024 State of the Global Workplace found fully remote employees reporting loneliness a lot of the previous day at 25%, hybrid at 21%, and fully on-site at 16%, with elevated rates under 35. That is a channel gradient measured with a subjective daily-experience item.
A February 2024 cross-sectional study of healthcare workers took the opposite approach and treated workplace isolation as a construct distinct from loneliness, with different correlates, finding that perceived social support moderated the relationship between remote working and well-being. Both are credible. They are not measuring the same thing, and a return-to-office policy justified by the first would not be justified by the second.
What a better instrument would record
The clinical recommendations now on the table require more precision than the instruments supply. The National Academies called for routine assessment of isolation and loneliness in health care, and the 2020 clinician-facing commentary in the American Journal of Geriatric Psychiatry worked through what that would demand in practice. The American Heart Association’s 2022 statement — which put the association with heart attack, stroke, or death from either at roughly 30% — identified absent intervention evidence as the field’s central gap. The HEAL-HOA dual randomised trial of volunteering among lonely older adults in Hong Kong, published this month in The Lancet Healthy Longevity, is notable mainly for being a randomised design in a literature dominated by uncontrolled evaluations.
The tractable fix is a channel supplement: alongside the existing subjective items, record for each contact whether it was in person, synchronous-mediated, or asynchronous, in the same respondents. American Time Use Survey evidence, discussed on WBUR in February 2024, already shows face-to-face socialising declining before the pandemic and continuing to fall after it — but time-use diaries and loneliness scales are almost never fielded on the same people. Until they are, the question of whether mediated contact substitutes for co-presence is not being tested. It is being assumed, in one direction or the other, by whoever chose the instrument.
Sources
- A Connected Society: A Strategy for Tackling Loneliness
- Loneliness and Social Connections: A National Survey of Adults 45 and Older
- Loneliness and the Workplace: 2020 U.S. Report
- Loneliness in America: How the Pandemic Has Deepened an Epidemic of Loneliness
- Almost a Quarter of the World Feels Lonely
- The State of Social Connections
- How Strong Are the World's Social Connections?
- Over 1 in 5 People Worldwide Feel Lonely a Lot
- 1 in 5 Employees Worldwide Feel Lonely
- Loneliness, Lack of Social and Emotional Support, and Mental Health Issues -- United States, 2022
- Loneliness and Social Isolation as Risk Factors for Mortality: A Meta-Analytic Review
- Social Isolation and Loneliness in Older Adults: Opportunities for the Health Care System
- Social Isolation and Loneliness in Older Adults: Review and Commentary of a National Academies Report
- Loneliness: Human Nature and the Need for Social Connection
- The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness
- The Evidence: Collaborative Review Documents on Adolescent Mental Health and Social Media
- Americans Don't Socialize Face-to-Face as Much as They Used To
- A Cross-Sectional Investigation on Remote Working, Loneliness, Workplace Isolation, Well-Being and Perceived Social Support in Healthcare Workers
- Effects of Objective and Perceived Social Isolation on Cardiovascular and Brain Health: A Scientific Statement From the American Heart Association
- Our Epidemic of Loneliness and Isolation: The U.S. Surgeon General Advisory on the Healing Effects of Social Connection and Community
- The State of Loneliness and Social Isolation Research: Current Knowledge and Future Directions
- The Effects of Volunteering on Loneliness Among Lonely Older Adults: The HEAL-HOA Dual Randomised Controlled Trial
- Bowling Alone: The Collapse and Revival of American Community