Why the Idea of a Loneliness Epidemic Won't Die

I have asked several colleagues why they keep calling loneliness an epidemic. More than one gave me the same answer. They know the word is not quite right, but it gets the topic into newspapers and onto ministers' desks. I understand the temptation, because the word works. In 2023 the U.S. Surgeon General titled his advisory Our Epidemic of Loneliness and Isolation. But a word also tells people where to look for a fix, and this one points the wrong way.

What an Epidemic Is

In public health, an epidemic is an increase in cases above what is normally expected in a population, often a sudden one. The word suggests something new with an agent you can find and remove, the way cholera led John Snow to the Broad Street pump in 1854. Loneliness has been measured for decades, so the question is whether it has risen suddenly.

What the Trends Show

For older adults, mostly not. Among Americans over 50, the share reporting loneliness fell from 20.1% to 15.5% between 1996 and 2018. Younger groups show more change. Loneliness in emerging adults rose slowly from 1976 to 2019 and stayed flat after 2012. Midlife loneliness crept up across generations in the United States and parts of Europe, while it held steady or fell in Nordic and continental Europe. A slow drift in some places and stability in others is not an epidemic.

What Our Reanalysis of Isolation Found

Loneliness is a feeling. Social isolation is an objective state, often measured as time spent alone, and plenty of people who spend time alone do not feel lonely. The advisory treated the two together and also cited a 2023 study of the American Time Use Survey that reported rising time alone from 2003 to 2020, with a sharp acceleration after 2018. When the authors declined to share their code, we organized six independent teams to reproduce the analysis, which I summarized here last year. Leaving out the pandemic year halved the apparent rise. What remained was about 24 more minutes alone per day over 17 years, a small effect (Cohen's d = 0.10). In follow-up work that is not yet published, Federico D'Atri and Randy McCarthy traced much of even this rise to who was surveyed. Over the years the data included more older adults, who spend more time alone. Differences between groups were far larger. Adults over 65 spent 231 more minutes a day alone than adults aged 25 to 34, nearly ten times the change over time. People without paid work spent 144 more minutes alone than those working 50 hours or more, and people with household incomes under $25,000 spent 98 more minutes alone than those earning over $100,000.

What the Label Does to Solutions

An epidemic calls for a new agent, and for loneliness the usual suspect is the smartphone. In three datasets covering more than 350,000 adolescents, digital technology use explained at most 0.4% of the variation in well-being. A cause that small cannot carry the epidemic story.

The larger factors are older. Middle-aged adults in Denmark, the Netherlands, and Sweden report the lowest loneliness of 14 countries studied, and the authors point to strong safety nets. Low income raises loneliness in later life, less so where welfare spending is generous. In our interviews in Türkiye, people kept coming back to money. Income, working hours, housing, and public space have no villain and no quick fix, so they make poor headlines. My colleague Neil Lewis Jr pointed me to Winners Take All (2018) by Anand Giridharadas, who described a similar pull, where funders favor solutions that leave the systems behind a problem intact. When the epidemic frame points at screens, budgets follow, and the larger levers go unused.

A Warning From Education

When we discussed these kinds of topics, Neil and another colleague, Andrew Przybylski, pointed me to education that has lived through this cycle. Controlled tests never supported learning styles, yet belief in them among educators remains high. Growth mindset was sold as a game changer, and a large meta-analysis found small average effects on achievement. Teachers in training who read research contradicting their beliefs tended to doubt the research itself, and learning-styles researchers warn that hype undermines trust in education research as a whole. Loneliness research is on the same road. If we promise an epidemic and a fix that turns out weak, city officials, employers, and clinicians will discount the next findings, including the ones that hold up.

Saying How Ready the Evidence Is

In 2020, colleagues and I proposed evidence readiness levels, modeled on NASA's technology readiness levels. The scale runs from early findings that need refining, through tests in real settings, to evidence ready for large-scale rollout. Most findings in our field had not reached the top.

The link between social connection and health rests on decades of long-term data. Claims that an app, a workplace program, or a screen-time limit will reduce loneliness at scale usually rest on far less. Naming the level lets a mayor or an HR leader judge how much weight a claim can bear.

There is a frame that wins attention without overstating the trend: Brain capital treats people's brain health and skills as an economic asset, and social connection is one of its drivers. An asset invites questions about investment and return, and those questions lead back to income, working hours, and public space.

Colleagues who use the epidemic label for attention could use the same platform to report the rate, the trend, and who is most affected, and say how ready the evidence for each fix is. If you fund programs, ask for that before you sign.

References