What a Social Connection Number Hides

In 2022, two large European surveys asked people the same loneliness question, in nationally representative samples from all 27 EU member states. They still produced very different prevalence rates and ranked the countries differently.

Governments now spend real money on loneliness, and they need tools that show whether a program changed what it set out to change. A loneliness figure can show the size of a problem but not what is missing in people's lives. One person lacks someone to rely on in a crisis. Another sees friends every week and still feels nobody understands them. Transport, housing, social services, and culture budgets each act on a different part of that picture.

Our latest policy brief for LONELY-EU, a Horizon Europe project, asks where the tools behind these numbers came from and what they actually ask.

What We Already Knew About Loneliness

With colleagues at the European Commission's Joint Research Centre, we tested common loneliness measures in 25,646 people across all 27 member states. Only the Three-Item UCLA Loneliness Scale produced scores that could be compared from one country to the next, and its scores track emotional states more closely than social activities.

Our previous policy brief drew the first lesson: A good questionnaire cannot give Europe a dependable picture on its own; it needs coordinated yearly data collection, quality checks, and a place in Eurostat's official statistics. Another policy brief then rated common measures green, yellow, or red on comparability across countries; whether they capture loneliness and not depression; and how fully they cover what people experience. None earned green on all three. I wrote about both on this page. Our new brief looks at the whole field.

Who the Tools Were Built For

This new policy brief draws on a systematic review Annecy Behavioral Science Lab ran with Ruhr University Bochum. We screened 9,032 articles, found 2,720 different tools, and focused on the 1,859 articles that developed a new one. Development is where the basic choices get made. A later translation can test a tool in a new country, but it cannot add questions the original authors never thought to ask. The United States accounts for 33.6% of development work, and English-speaking countries for 56%. Only 18.7% involves an EU member state. Among tools with a clear age band, 52% center on adults aged 26 to 44 and 4.7% on adults aged 65 and older. Only 10.8% mention an ethnic or racial group at all.

What the Tools Ask

Researchers often sort social connection into structure, function, and quality: how many relationships people have and how often they see others, what those relationships provide, and how good or strained they feel. We coded every question in all 2,720 tools into 73 content categories, so any two tools could be compared. Tools in the same part overlap far less than their labels suggest. Two tools that both measure function share on average 15.6% of their content, structure tools 20.1%, and quality tools 9.4%. Depression scales, a field already known for varied measures, share around 36%. Two figures with the same label can describe different experiences. The survey matters too, as mode, sampling, and question order also shape results.

Different Places Need Different Solutions

The low overlap reflects how varied social connection is. In our interviews across countries, people describe connection in specific terms, such as someone to rely on, feeling understood, trust, or safety, and which of these matters differs from person to person and place to place.

For policy, that means a program that works in one setting can miss in another. A befriending scheme may help someone who lacks contact but do little for someone surrounded by family who feels misunderstood. Governments need monitoring that shows which form of connection is missing, for whom, and where, so each problem goes to the department that can act on it. Measurement has to be built for that variety, with tools developed for the people being measured.

The same gap emerges in the newer economic language around the issue. Brain capital treats people's brain health and skills as an asset worth investing in, and counts social connection as one of the drivers. So far that driver is tracked with only a couple of broad indicators. If governments are going to invest in connection the way they invest in skills, they need to know which form of it they are buying, and for whom.

What Governments Can Do Now

A long series that keeps the same question and method is still the best evidence of change over time. The brief asks statistical offices to keep those series and to publish, next to each figure, the tool used, what it captures, and how the survey was run. For new work, it recommends pairing the Three-Item UCLA scale with a few questions on contact and on having someone to rely on. Funders can ask that new tools be developed with older adults, migrants, minorities, and people in economic precarity from the start.

Within LONELY-EU we are building the EU Social Isolation and Loneliness Index for this purpose. Three core items keep countries comparable, 10 extended items serve national monitoring, and 10 more can vary between groups such as older adults. The extended version reports each form of connection separately, so a ministry could see whether a region's gap lies in contact, in support, or in feeling understood. Like any new tool, it needs independent validation before it can carry policy weight.

If you work with these numbers, look up the tool behind the next statistic you cite in the LONELY-EU Item Explorer, which shows what each published tool asks. Then check whether those questions match what your program is trying to change.

References