A survey of 1,408 UK students found loneliness almost everywhere it looked

A cross-sectional survey of UK university students published in npj Mental Health Research on September 16 reports a prevalence figure that is striking and, on inspection, difficult to use: moderate to severe loneliness in 78.98% of 1,408 respondents [s1].

The paper's own abstract describes that figure as "approximately two thirds" of students [s1] — an internal inconsistency in the published text that a reader should notice before quoting either number.

What was measured

Participants completed eight validated instruments: the Generalized Anxiety Disorder scale, the Patient Health Questionnaire, the Mood Disorder Questionnaire, the Sleep Condition Indicator, the Perceived Stress Scale, the Suicidal Behaviours Questionnaire–Revised, the Prodromal Questionnaire 16, and the UCLA Loneliness Scale [s1].

That is a substantial battery, and it allows the study's more interesting analysis: not just whether loneliness correlates with each symptom, but which correlations survive once the symptoms' shared variance is accounted for.

The findings

Loneliness was significantly associated with every psychiatric symptom measured [s1].

After accounting for shared variance between symptoms, linear regression found loneliness remained significantly related to perceived stress, depression, psychotic-like experiences, and suicidal ideation [s1].

The inclusion of psychotic-like experiences in that surviving set is the least expected result. The association between loneliness and depression is thoroughly documented; a relationship with prodromal psychosis symptoms that persists after adjusting for depression and anxiety is less so, and the study flags it as a prominent finding [s1].

The problem with a prevalence of 79%

A measure that classifies roughly four in five members of a population as affected is doing something other than identifying a distinct group. Either loneliness at moderate severity is close to the normal condition of being a UK university student — which is possible, and would itself be worth knowing — or the threshold applied to the UCLA scale is set where it separates very little.

This matters for the correlational findings too. When an exposure is nearly universal, the comparison group shrinks and becomes unrepresentative, and associations between the exposure and other symptoms become harder to interpret.

What the design cannot support

This is a cross-sectional survey [s1]. Every association reported is a snapshot. Loneliness may contribute to depression, depression may cause withdrawal that produces loneliness, or a third factor may drive both; the data cannot separate these, and the study does not claim to.

Recruitment method is central to interpreting a prevalence figure of any kind, and a self-selected sample of students willing to complete eight questionnaires about mental health will not be representative of students generally — most plausibly, in the direction of over-representing those with symptoms.

All eight instruments were self-report, completed in the same sitting. Correlations between self-report measures administered together are inflated by shared method variance, which the adjustment for shared symptom variance addresses only partly.

Where it sits in a larger picture

A cross-sectional analysis of Gallup World Poll data published two weeks earlier found that global social isolation — measured as having no relatives or friends available to help in times of trouble — rose 13.4% between 2009 and 2024, from 19.2% to 21.8%, with the entire increase occurring after 2019 [s2].

The two studies measure different things, and the gap between 21.8% and 78.98% is largely that difference rather than a difference between populations. Isolation as a missing practical resource is rarer than loneliness as a subjective state, and the distinction is not cosmetic: interventions that expand someone's network address the first, while the second can persist inside a full social calendar.

The student study measures the subjective construct, in a population that has social contact available almost by definition. That is precisely why its prevalence figure is high — and why "tailored interventions for lonely students," which the authors call for [s1], will need to be specific about which of the two problems they are treating.

What to watch

Whether longitudinal student cohorts show loneliness preceding depressive and psychotic-like symptoms rather than accompanying them; whether the UCLA scale threshold used here is applied consistently across the student literature; and whether the psychotic-like experience association replicates in samples with clinical follow-up.

This article describes a cross-sectional survey and is informational only. It is not medical advice.

If you or someone you know is struggling, contact a local crisis line or health service.

Sources

• [s1] Akram U, Drabble J, Irvine K, Allen S, Akram A, Stevenson JC, Gardani M, Prevalence and psychiatric correlates of loneliness in UK university students, npj Mental Health Research, 2025;4:45, published 2025-09-16.
• [s2] Fuller-Rowell TE, Sultana S, Kawachi I, Global Trends and Disparities in Social Isolation, JAMA Network Open, 2025;8:e2532008, published 2025-09-02.