What is Snowball sampling?
A non‑probability recruitment method where existing study participants recruit future participants from their social networks, causing the sample to grow like a snowball. It's often used to reach hard‑to‑find or networked populations.
Snowball sampling starts with a small group of initial participants (seeds). Each seed is asked to refer people they know who meet the study criteria; those new participants in turn refer others, and the sample expands through social connections. Because participants recruit people they already know, the method is quick and cost‑effective for reaching hidden, small or tightly connected groups, but it does not produce a statistically representative sample and can introduce bias.
Usage example
A local council needs feedback from recently arrived refugee families who speak a range of languages. The engagement officer asks a few community leaders to complete the survey and invite other families they know to take part. Each respondent passes the survey link to friends and relatives, rapidly increasing participation among that community.
Practical application
Why it matters: Snowball sampling is useful when you need to reach people who are hard to contact through standard channels (e.g., recent migrants, people experiencing homelessness, niche online communities) and when time or budget for recruitment is limited. In Hearo's multilingual context it can help surface responses from language communities that would otherwise be missed. Trade‑offs: results are prone to network bias (over‑representing certain social circles) and should not be used to make population‑level prevalence estimates without caution. Mitigations include using multiple, diverse seeds, limiting referrals per participant, tracking referral chains, collecting demographic data to understand coverage, and combining snowballing with other recruitment strategies.
FAQ
Is snowball sampling representative of the whole population?
No. Because participants recruit from their own networks, some groups can be over‑ or under‑represented. Snowball samples are useful for exploratory and qualitative work or to reach hidden groups, but not for producing precise population estimates without careful adjustment and transparent limitations.
How is snowball sampling different from convenience sampling?
Both are non‑probability methods, but convenience sampling recruits whoever is easiest to reach. Snowball sampling uses participant referrals to reach others in similar networks, which can be better at finding people who are otherwise inaccessible but also reinforces network biases.
How can I reduce bias when using snowball sampling?
Start with multiple, diverse seeds (different ages, neighbourhoods, language groups), limit how many referrals each participant can make, track who referred whom, and collect basic demographic data to check which groups are missing. Where possible, combine snowball recruitment with other outreach (community organisations, targeted advertising) to broaden coverage.
Are there ethical issues with snowball sampling?
Yes. Privacy and consent are important because referrals involve sharing contact details or links. Make it clear to participants how referrals will be used, avoid asking people to share sensitive data about others, and provide an easy way for invitees to opt out. Be sensitive when recruiting from vulnerable communities and follow relevant safeguarding and data‑protection guidance.