Snowball sampling recruits further participants through referrals from people already contacted. It can help researchers reach a group for which a practical contact list is unavailable. It is a non-probability method: social connections and willingness to participate shape who enters the study.
Worked example: students returning after a study break
Imagine a small interview project about returning to university after a long break. An approved student-support notice recruits three eligible volunteers. At the end of their interviews, the researcher asks whether they would share the public invitation with other eligible students. Those students contact the researcher themselves if interested.
This is an original planning scenario, not a completed study. The initial volunteers are starting points for referral chains. Their contacts may share courses, friendships or experiences, so the final group can reflect those networks more strongly than the wider population.
A recruitment plan you can explain
| Decision | Example plan | Why record it? |
|---|---|---|
| Eligibility | Currently enrolled students returning after a defined break | Shows who could participate |
| Starting points | Invitations through several approved support channels | Makes network dependence visible |
| Referral process | Participants share an invitation; no contact details are collected about friends | Protects voluntary contact |
| Stopping | A justified project limit reviewed with the supervisor | Separates practical limits from claims of analytic adequacy |
Keep consent separate from friendship
Agree the process with the relevant ethics reviewer before recruitment. An existing participant should not promise someone else will join or reveal sensitive information about them. A shareable invitation can let interested people approach the researcher privately. Explain participation, withdrawal arrangements and how data will be handled using the actual approved study documents.
If the project involves a particularly sensitive or identifiable group, even an invitation may carry risk. Discuss those details with the supervisor rather than treating this example as automatic ethics approval.
What the findings can support
Suppose eight interviewees describe difficulty reconnecting with classmates. You can analyse how those interviewees describe the problem and compare their accounts. You cannot infer the percentage of all returning students experiencing it from this referral sample.
Seek different starting points when useful to the question, record which channels worked and report gaps. These steps can broaden the experiences encountered, but do not turn referrals into random selection. Do not claim representativeness just because the referral chain became long.
Example methods wording
Initial participants responded to approved student-support invitations. They could share the study invitation with other eligible students, who contacted the researcher voluntarily. Referral recruitment supported access within the project timetable, but may have favoured connected students and similar experiences. Findings describe the participating group rather than population prevalence.
Replace this wording with your actual eligibility criteria, channels, recruitment dates, achieved sample and limitations. Describe any purposive criteria alongside the referral process; they answer different questions about selection.
Before you start
- Define eligibility and explain the referral route.
- Use consent materials approved for your project.
- Consider who is disconnected from the starting networks.
- Record recruitment changes and unsuccessful channels.
- Keep claims within the design and evidence.
Method background: Statistics Canada: non-probability sampling. The university scenario and methods paragraph are original teaching examples.
Related: purposive sampling, convenience sampling and choosing a sampling method.