What actually works for making friends as an adult?

What has been tried?

Researchers group loneliness interventions into four kinds. Improving social skills, enhancing social support, increasing opportunities for contact, and addressing maladaptive social cognition, which means the beliefs and expectations a person brings to social situations.

The first three are what most advice recommends: join a club, take a class, meet more people. The fourth is the least intuitive and it is the one the evidence favours.

Which approach works best?

Changing how you interpret social situations. In a meta-analysis of 50 studies, interventions addressing maladaptive social cognition outperformed the others among randomised comparisons. The overall effect in randomised studies was modest at around −0.20.

Masi, Chen, Hawkley and Cacioppo published this in Personality and Social Psychology Review. Maladaptive social cognition covers the assumptions that quietly shape a social encounter before it happens: expecting rejection, reading neutral responses as coldness, assuming an unanswered message means something.

That is a plausible mechanism rather than a comfortable one. Loneliness appears to make people more alert to social threat, which produces guardedness, which produces the distance that confirms the expectation.

Why is the effect so modest?

Partly because loneliness is difficult to shift, and partly because most of the encouraging numbers come from weaker designs. Single-group before-and-after studies produced a mean effect of −0.367 and non-randomised comparisons −0.459, against −0.198 in randomised trials.

That gap is worth naming plainly. Effect sizes roughly halve when a proper control group is added, which means most of the figures quoted for loneliness programmes are inflated by design rather than by outcome.

The authors also flag heterogeneous study quality, limited follow-up on whether gains last, and the finding that studies using one particular loneliness scale reported larger effects.

So what should you actually do?

Start with the expectations rather than the calendar. Notice what you assume when a message goes unanswered or an invitation is declined, and check whether the harshest interpretation is actually the most likely one. That is the lever with the best evidence behind it.

Then use repetition rather than events. Friendship in adulthood is built by turning up to the same thing repeatedly with the same people, which is why classes, teams and regular groups work better than one-off gatherings. Structure does the work that spontaneity did when you were younger.

Be the one who initiates, and expect it to feel disproportionate. People consistently underestimate how much others enjoy hearing from them, so the message you are hesitating over is usually more welcome than it feels.

Group formats and interventions that create repeated interaction both showed potential in this literature, so this advice sits alongside the evidence rather than against it.

Is loneliness the same as being alone?

No, and the distinction matters for what to do. Loneliness is the gap between the connection you have and the connection you want, which is why it can appear in a full social calendar and be absent in a quiet life.

This connects to the mortality evidence on this site, where objective isolation carried more weight than the subjective feeling. Both matter, and they call for different responses: circumstances for one, interpretation for the other.