Group therapy has a substantial evidence base — meta-analyses generally find it comparable to individual therapy across a range of conditions, and for some presentations better. The interesting question isn’t whether it works but by what mechanism, because the mechanisms tell you which groups suit which people.
Yalom’s list of therapeutic factors is the standard reference. Here are the ones that do most of the work.
Universality
The discovery that the thing you were certain was uniquely and shamefully yours is sitting in the room with three other people.
This is the most consistently reported effect and it’s the one people describe as changing something immediately. A therapist telling you that many people feel this is information. Hearing a person describe your exact private experience in their own words is a different order of event, and it reaches the part of you that had decided you were the exception.
It’s also why group is so effective for shame specifically —and for social anxiety, where the conviction of being uniquely defective is the disorder.
Interpersonal learning
The mechanism that individual therapy can’t reproduce.
Whatever you do with people, you’ll do in the group. If you withdraw when you feel criticised, you’ll withdraw in there. If you make jokes when things get serious, you will. If you look after everyone else so that nobody looks at you, that will happen too.
And then — this is the part — people tell you. Not your therapist’s interpretation of your account, but eight people’s direct experience of being in a room with you. That feedback is almost impossible to obtain anywhere else in adult life, because the people who could give it are either too polite or too involved. It’s also the thing people flinch from most.
Cohesion
The group equivalent of the therapeutic alliance, and the single strongest predictor of outcome in the group research. Belonging to something that meets weekly and notices when you’re absent does a substantial amount on its own, particularly for loneliness.
Altruism
Being useful to someone else. Underrated, and frequently the first thing people report as surprising — that helping another member did more for them than being helped did.
For anyone whose difficulty involves feeling like a burden, this one isn’t a side effect. It’s the treatment.
Vicarious learning
Watching someone two months ahead of you do the hard thing. Seeing that it’s survivable when someone else does it lowers the bar considerably more than being told it’s survivable.
The corrective family experience
Groups tend to reproduce something of the family you grew up in — rivalry, favouritism, who gets the attention, what you’ve to do to be noticed. Which sounds like a problem and is the opportunity: it’s the same dynamic, in a room where it can be named out loud and responded to differently.
What this implies practically
If your difficulty is interpersonal — how you’re with people —a process group is where those mechanisms are strongest.
If you need to learn a specific set of skills, a skills group works through a narrower set of mechanisms and works fine.
And if what you need is other people who share one specific situation, a support group delivers universality and cohesion without the rest, which for grief and for illness is frequently exactly the right amount.