Recruiting the People Who Won't Sign Up

A poster in the GP waiting room. A flyer at the library. A post on the community Facebook group. This is how most programmes recruit, and it works — for people who read posters, visit libraries, and follow community Facebook groups. It systematically fails to reach the person it was probably designed for: someone with a shrinking network, no habit of showing up to things, and no one nearby to mention the poster to.

This is not a minor gap in reach. It is the central design problem of the field. If your recruitment channel depends on someone already being somewhat socially connected — checking a noticeboard, scrolling a group, having a friend who forwards a link — you have built a filter that removes exactly the people most affected by isolation before they ever hear about the programme.

Why standard recruitment filters out the target population

The predictors of loneliness are not random. The AARP Foundation’s 2018 national survey of adults 45 and older found the strongest predictors were the size and diversity of a person’s social network and their degree of physical isolation — and the gap is stark: 33% of people who talk to their neighbours report loneliness, against 61% of those who never have. Robert Putnam’s account of declining civic participation describes the same mechanism at a societal scale: the associational structures that used to pull people into rooms — unions, congregations, bowling leagues — have thinned out, and thinning networks don’t recruit new members into replacement activities either. The 2021 survey from the Survey Center on American Life found 12% of Americans report no close friends at all, up from 3% in 1990, and 15% of men report the same. A flyer strategy assumes a friend to pass it to. For a growing share of the population, there isn’t one.

KFF’s 2024 survey adds another layer: loneliness and social support vary by race and ethnicity in ways most national surveys don’t capture, and discrimination experiences are tied to smaller support networks. A recruitment channel that runs through existing community infrastructure will reproduce whatever exclusions already exist in that infrastructure.

The failure mode: the volunteer-shaped hole

Call it the volunteer-shaped hole. Programmes designed to combat isolation get filled, disproportionately, by people who are not very isolated — the retired teacher who joins everything, the newcomer who is confident and just needs local knowledge. They benefit, they say kind things in the evaluation, and the funder sees healthy attendance. Meanwhile the person the programme was actually funded to reach never appears, because reaching them was never actually built into the plan — only hoped for.

This matters for evaluation as much as for ethics. A programme that reports strong satisfaction scores from a self-selected, already-connected group is not evidence the model works for isolated people. It’s evidence it works for people inclined to try new things.

Four channels that reach past the self-selecting

None of these are exotic. They are simply the ones that don’t rely on the target person already having initiative and a network.

1. Route through people who already have mandated contact with them. GPs, pharmacists, home care workers, housing officers, meals-on-wheels drivers, probation officers. Social prescribing depends entirely on this: the link worker model works because it inserts a referral point into a contact the person didn’t have to seek out. The qualitative synthesis on social prescribing found that participants valued benefit that went beyond social contact itself — restored purpose and participation — but that value only reaches people once someone with existing contact makes the referral. If your programme has no relationship with primary care, home care agencies, or housing services in its catchment, it has no route to the people who never leave the house voluntarily to look for one.

2. Recruit through routine, not through choice. A shopping trip, a pharmacy collection, a laundromat, a community meal already on someone’s calendar for other reasons. Anywhere a person turns up without deciding to “get involved” is a better recruitment site than anywhere requiring a decision to join something new.

3. Use warm handoff, not information. A leaflet is information. A staff member walking someone to the first session, or a phone call rather than a mailed invitation, is a handoff. The evidence on isolated older adults suggests network size and physical isolation are the barrier — an isolated person is, by definition, less likely to convert information into action alone.

4. Build in a no-cost, no-commitment first contact. The randomised trial of a volunteering intervention among lonely older adults in Hong Kong — one of very few controlled trials in this literature — tested prosocial engagement against a genuine control group, which is itself instructive: most loneliness interventions are evaluated without one, so it’s rarely clear whether recruitment channel or intervention content is doing the work. Where a low-stakes taster exists, it should be the entry point, not a twelve-week sign-up.

What this means in practice: if your recruitment plan is a list of places to advertise, it is not a recruitment plan for the hardest-to-reach group — it’s a recruitment plan for the easiest one. Build the plan around existing points of mandated or routine contact first, and treat posters and social media as a supplementary channel for people who were always going to find you.

Evidence status of the four channels

Channel	Evidence status
Referral through health/care contact points	Reasonably supported qualitatively; social prescribing literature consistently shows value once referred, less evidence on referral completion rates
Recruitment via routine (non-programme) contact	Logical extension of network/isolation research; not directly tested as a recruitment method
Warm handoff vs. leafleting	Plausible given isolation’s relational nature; no controlled comparison exists
Low-stakes taster as entry point	One relevant RCT (volunteering, Hong Kong); not a recruitment study, and results shouldn’t be generalised beyond that population

What this does not solve

Even a referral network through GPs and home care misses people who have no contact with any service at all — no GP registration, no benefits claim, no housing officer. That population exists and is likely the most isolated of all. No recruitment channel described here reaches someone who has fallen out of every system simultaneously; that requires outreach workers physically finding people, which is a resourcing decision most programme budgets don’t make room for. Naming the four channels above is a floor, not a ceiling, and even a well-built referral network will still undercount the people with no institutional contact to route through.