Snowball sampling
Reach participants through social connections
Snowball sampling recruits through referrals from initial participants or contacts. It is often non-probability sampling in practice and useful where a conventional frame is missing. Goodman's formal network design uses random initial selection and specified referral stages; informal snowballing generally lacks those conditions. Respondent-driven sampling adds its own procedures and assumptions and should not be treated as synonymous.
Use to reach connected or difficult-to-identify populations when trusted introductions improve access, while protecting confidentiality and actively checking the diversity of recruitment chains.
Strengths
- Enables access beyond formal institutional lists
- Trusted referrals can support initial contact
Limitations
- Highly connected people and seed networks may dominate
- Referral relationships create privacy and coercion concerns
Know the boundary
Informal chain referral does not inherit Goodman's probability estimators or respondent-driven sampling assumptions.