Craft, Making and Mental Health: Knitting, Pottery and Men's Sheds

Community making, meaning knitting groups, pottery, woodwork, repair cafes, allotments and men’s sheds, is social infrastructure that happens to involve craft, it is not therapy, it does not claim to be, and it does not need to be in order to be worth funding. The claims made on its behalf, mostly by people who love it and want it to survive, are the one thing about it that needs handling carefully.

The men’s shed I know best is in the unit next to the room where I rehearse, in a small industrial estate with a shared car park and a fire door that neither of us can prop open in winter. I have been in and out of it for about four years, mostly to borrow things. It is a woodwork shop with two lathes, a bench for bike repairs, and a kettle that gets more use than either. On a Tuesday there might be fourteen people in it. Some of them are making something. Several are not.

The thing its organiser said to me, which I have repeated more often than anything else I have picked up in this sector, came out of a conversation about a funding application. A local health partnership had money available and he was thinking about applying, and I asked whether he would have to change anything to qualify. He said the money was fine but he would not put the word wellbeing on the door, and then he said: “If I called this a wellbeing group, half of them would stop coming and the other half would stop talking. They come because there’s a lathe. What happens after that is nobody’s business but theirs.” He has run the place for eleven years and he knows exactly who walks past the door and who comes in. I think he is right, and I think he is right in a way that the sector’s own funding language finds genuinely difficult to accommodate.

Where a claim about effect appears below it carries a tier: Tier 1 supported, Tier 2 promising but limited, Tier 3 practice and anecdote. What each label requires is set out on reading arts in health research.

What the field actually covers

Community making covers knitting and stitching groups, pottery and ceramics, woodwork and metalwork, repair cafes and bike workshops, allotment and growing projects, and the men’s shed format that sits somewhere between all of them. The unifying feature is not the craft. It is that people make things in the same room as other people, regularly, over a long period.

That distinction is worth taking seriously rather than treating as a nice observation, because it explains almost everything else about these groups. Most of them are social infrastructure first and craft second. The knitting group in a library where nobody has finished a jumper in two years is not failing at knitting; the knitting is the reason to be there and the ostensible subject of conversation, and it is doing its job perfectly. A repair cafe that fixes eleven toasters in an afternoon is also a place where four people who live alone have talked to somebody.

The formats vary in ways that matter to who attends. Some are open access and free, some charge a small subscription, some run in blocks of weeks with a waiting list. Some are attached to a service, a hospital, a hospice or a care home, and some are entirely independent of anything, which is often their principal asset. Some are single sex by design and some are not, and the argument about that is a real one rather than an oversight.

They also differ from the arts side of this field in one practical respect. A making group produces objects that leave the room and get used, given away or sold, and that changes the relationship between the person and what they make. Nobody hangs a repaired kettle on the wall.

Why people say it helps

Here is the account people give of why these groups work, and it is presented here as a plausible mechanism rather than as evidence, because that is precisely what it is.

Occupied hands. A task takes up the part of attention that would otherwise be spent monitoring the conversation and monitoring yourself. People say things over a lathe that they would not say across a table, and they say them without the weight that comes from having sat down specifically in order to say them.

A low demand on conversation. Silence in a workshop is not awkward, because the noise of the room fills it and everybody has something to be doing. A person can attend for six weeks, say almost nothing, and still be a member of the group in a way that would be impossible in a circle of chairs.

Sitting side by side rather than face to face. This one comes up so consistently, from organisers of very different kinds of group, that it is worth naming on its own. Facing somebody is an interrogative posture. Standing next to them at a bench is not.

A task with a visible end. Depression and anxiety both eat the sense that effort leads anywhere. A finished bowl, a mended chair, a row of stitches that was not there this morning is a small piece of counter evidence that can be held in the hand.

Somewhere to be on a Tuesday. The least glamorous item on the list and probably the most important. A fixed point in the week that expects you, notices if you are not there, and does not require you to be well in order to attend.

Every one of those is a story about a mechanism. They are coherent, they match what participants describe, and they are the reason I keep turning up to my own sessions. None of them is a finding. A mechanism that sounds right is the easiest thing in the world to believe and the hardest thing to test, and the gap between “this is why it would work” and “this works” is where most of the overclaiming in this field is manufactured.

Men’s sheds, and the word on the door

Men’s sheds reach a group of people that health branded provision reliably fails to reach, and the reason is almost entirely about framing rather than about content.

The format is straightforward: a workshop, tools, set opening days, a subscription of a few pounds, and members who run the place themselves rather than being run by a service. Many are open to everybody despite the name. What they offer is a task and a reason to be in a building, and what they conspicuously do not offer is an invitation to discuss how you are.

The organiser’s line about the word wellbeing is the whole argument in miniature, and I want to let him be right rather than translating him into something more comfortable. For a substantial number of people, mostly but not only men, a group described in terms of mental health is a group that announces something about you before you have taken your coat off. Attending it requires you to have accepted a description of yourself that you may not have accepted, and to accept it in public, in front of neighbours. Attending a woodwork session requires nothing at all. You are there because there is a lathe.

That is why the funding conversation is genuinely hard rather than merely bureaucratic. The money available for this kind of provision often comes through health routes and expects health language, health outcomes and health reporting. The groups that would benefit most from it are frequently the ones whose effectiveness depends on not using any of that language. Some organisers take the money and quietly keep the door sign unchanged. Some do not apply. Neither is a satisfactory answer, and pretending the tension is not there helps nobody.

It is not therapy, does not claim to be, and does not need to be

None of this is therapy in the regulated sense, and the useful thing about saying so is what it protects rather than what it withholds.

In the UK four arts therapy titles are protected in law: art psychotherapist, art therapist, dramatherapist and music therapist, and using any of them without registration is a criminal offence. In July 2026 the Health and Care Professions Council register listed 6,103 arts therapists across all four titles¹. A registered arts therapist brings an assessment, written goals, clinical notes, clinical supervision, accountability to a regulator that can strike them off, and a planned ending. A craft group brings none of that.

Three things follow, and they are the same three that the sibling pages on this site keep making, because they are the field’s central confusion.

Arrangement does not transfer. Therapy is arranged by referral, assessment and clinical decision. A making group is arranged by turning up, or by a link worker’s suggestion, and most take self referral. These are not two doors into the same building.

Regulation does not transfer. The protection attaches to the title and to the person, not to the activity. Someone can lawfully run a superb therapeutic making group for twenty years; what they cannot do is call themselves an art therapist without being on the register.

Evidence does not transfer. A result proved about a clinical intervention delivered by a registered practitioner says nothing about a Tuesday workshop, and vice versa. This is the point the sector’s own literature loses most often, usually while quoting something in good faith.

This page is the fourth of four carrying that distinction. The others are art therapy against an art class, music therapy against community music and dance movement psychotherapy, and the regulated half as a whole is mapped on art therapy and art psychotherapy. The reason to insist on the line is not to demote the craft group. It is that describing an unregulated activity in clinical language sets it up to be judged against clinical evidence it was never built to produce, and it will lose that test every time.

What has actually been measured

The closest measured body of work is not about craft specifically. It is the arts on prescription literature, and the 2024 systematic review is the honest summary of what is in it.

That review screened 7,805 records and included 25 studies: 10 qualitative, 7 mixed methods and 8 quantitative uncontrolled before and after studies. Its own words, verbatim, are that “No Randomized Controlled Trials (RCTs) were identified in the search”. Most of the interventions ran 8 to 10 weeks². Its meta analysis found a statistically significant improvement in wellbeing.

That last sentence is the one that travels, so here is what sits underneath it. The pooled result rests entirely on uncontrolled designs, which means every one of the quantitative studies measured a group before a course and again afterwards, with nothing to compare against. Such a design cannot separate the activity from time passing, from regression to the mean, from who chose to sign up, or from who dropped out before the second measurement. So on this site the finding is Tier 3, and it is never called a trial finding, whatever the document quoting it calls it. That is not a criticism of the review, which is scrupulous about its own limitations; it is a description of the material the review had available.

Tier 3 here is the kind that means nobody has run the adequately powered controlled study, not the kind that means somebody ran one and it came out null. Those two situations get one label and mean opposite things, and this is the point at which readers most often draw the wrong conclusion. The wider arrangement of these schemes, and who refers into them, is described by the National Academy for Social Prescribing³, which is a description of practice rather than evidence of effect. See arts on prescription, social prescribing explained and what a ten week arts course involves for how the system works in practice, and does arts on prescription save money for the savings figures, which do not survive contact with their source.

It is also worth knowing what the field looks like from a distance, so that its size is not mistaken for its strength. The WHO Europe scoping review mapped over 900 publications, comprising 200 plus reviews and 700 plus individual studies, with those reviews between them covering over 3,000 studies⁴. The counts nest rather than add: the 3,000 sit inside the reviews, which sit inside the 900, and setting them side by side makes the literature look several times larger than it is. A scoping review maps a literature rather than pooling it, so it carries no tier and supports no claim about effect at all. It establishes that a great deal has been written. It says nothing about what was found.

How a claim inflates, in about three citations

Watching a claim grow is more useful than being warned about it, so here is the sequence, which I have now seen run to completion more than once.

Step one, the evaluation. A group runs for ten weeks. Twenty two people complete a questionnaire at the start and at the end, and the average score improves. The report says so accurately, notes that there was no comparison group, and includes some quotations from participants who found it valuable. Everything in it is true and carefully hedged.

Step two, the summary. Somebody writing a regional strategy needs a sentence. The hedges are the first thing to go, because a strategy document has no room for a paragraph about uncontrolled designs. The sentence becomes “participants reported improved wellbeing”, which is still true, and it now carries a footnote to the evaluation that most readers will not follow.

Step three, the citation of the summary. A funding bid quotes the strategy rather than the evaluation, because the strategy is the authoritative looking document and it is easier to find. “Participants reported improved wellbeing” becomes “craft groups improve wellbeing”. The subject of the sentence has quietly changed from the people to the activity, and a claim about what was reported has become a claim about what was caused.

Step four, the round number. Somebody adds a figure, usually a percentage taken from the original evaluation’s own before and after scores, sometimes from a different study altogether. Now there is a statistic. From here the claim is quotable, and within another citation or two it will be “reduces depression”, which nothing in the chain ever showed and which the original report explicitly did not claim.

Nobody in that chain is dishonest. Each step is a small, defensible compression made by somebody under time pressure who is trying to keep a valuable service open. The damage is done in aggregate, and it lands in two places: the field acquires an evidence base it does not have, and any commissioner who eventually checks the trail concludes that the whole sector overclaims, including the parts that do not. Interrupting the sequence at step two is the whole purpose of this site, and the method for doing it is on reading arts in health research.

How these groups are funded, and why they close

Community making is funded in short cycles from a patchwork of sources, and the closure of a good group usually has nothing to do with its quality.

The typical mix is a local authority grant, a National Lottery award, a charitable trust, occasionally a health and care partnership, a parish or town council, membership subscriptions, and the sale of whatever gets made. Cycles run one to three years. Almost none of it is recurrent, and most of it comes with reporting requirements written for a different kind of organisation.

Three predictable things follow. Groups spend a disproportionate share of their effort on applications rather than on the group. They are pushed towards describing themselves in whichever language this year’s fund prefers, which is where the wellbeing problem starts. And a well attended, well run group can close at the end of a term because a grant cycle ended, taking with it the one fixed point in a number of people’s weeks. Where the money comes from in more detail is on who pays for arts on prescription.

The organiser at the shed puts it more bluntly than any funder would. His subscription income covers the rent and nothing else, and the equipment came from three separate grants and a bequest. If the rent goes up he closes, and eleven years of accumulated knowledge about who needs a phone call when they miss two Tuesdays goes with it.

How to find one

Most making groups take self referral, which means no diagnosis, no referral letter and no assessment, and you can generally just appear.

The routes worth trying are a social prescribing link worker at a GP practice, the library, the community centre noticeboard, the local authority’s community directory, a parish magazine, and the men’s shed association listings for that format specifically. If a link worker is involved, how to get referred to an arts scheme sets out how the referral actually works, and it is a shorter process than most people expect.

Reasonable things to ask before you go: what it costs, whether there is a subscription or a waiting list, whether tools and materials are provided, whether you can come once and see, and what happens if you miss a few weeks. A group that answers those easily is a group that has thought about newcomers, which is the single best predictor I know of whether somebody will still be attending in three months.

The reason to defend these places is not that they are a cheap substitute for something clinical. It is that they are a good version of a different thing: a reliable, unpressured, unlabelled reason to leave the house, run by people who will notice when you stop coming. That is worth funding on its own terms and does not need a borrowed vocabulary to justify it. And if somebody is unwell and being treated for it, the workshop sits alongside that treatment and never in place of it; a lathe has never yet replaced a prescription, and no organiser I have met would suggest otherwise.