Does talking to strangers actually help with loneliness?

Partly, and not in the way the question usually means. Both halves of that are worth spelling out, because this is the page where it would be easiest to tell you what you want to hear.

What the evidence supports

Brief conversations with strangers reliably lift mood, and by more than people expect. That result comes from field experiments where commuters were randomly assigned to talk to somebody or not — random assignment, so this is a genuine causal claim rather than a correlation, and it has been reproduced in more than one country.

There is also a body of work on what psychologists call weak ties — the acquaintances, colleagues, baristas and classmates who make up most of the people you interact with. People report greater happiness and a stronger sense of belonging on days when they have more of those interactions than usual. It's a real and slightly surprising pattern, and it is the closest thing to evidence that these interactions do something beyond the hour they happen in.

What the evidence does not support

That talking to strangers cures loneliness. Nobody has shown that, and the studies are not designed to.

The mood research measured how a single commute felt. Not the week, not the month, not whether anybody was less lonely in April. The weak-tie research is largely correlational — people who have more of these interactions feel better, but you cannot tell from that whether the interactions lift the mood or the mood produces the interactions, and it is very likely both.

And loneliness in the sense people usually mean it — the absence of somebody who knows you — is a different thing from an under-supply of daily contact. A stranger cannot know you. Nothing in this literature suggests brief anonymous conversation substitutes for a close relationship, and a site that told you otherwise would be selling something.

So what is it actually good for

Three things, all modest. The first is measured; the second and third are us reading across from studies that were not asking this question:

• The evening itself. The mood effect is measured and reproducible. It is not life-changing and it is not nothing.
• Practice. This is the better argument. If avoidance is part of why the closer relationships aren't there, then repeated low-stakes conversation is the one thing shown to shift the pessimism behind it.
• Evidence about yourself. Most people carry a badly calibrated estimate of how they come across. Conversations are the only thing that corrects it, and the estimate is wrong in a consistent direction.

The honest summary

If you are lonely in the sense of having nobody close, this is not the answer, and we would rather say so than have you find out over six months. What it can be is the cheaper end of the same work: contact that reliably improves an evening, and repetition that appears to loosen the thing keeping you from the rest of it.

If loneliness is heavy enough that most days are hard, that is worth a doctor rather than a website. That is not a disclaimer to get out of the question; it is the actual answer to it.