Field-based Snowball Sampling
Also known as: in-person snowball sampling, fieldwork chain-referral sampling, field snowball sampling, face-to-face referral sampling
Field-based snowball sampling is a non-probability chain-referral technique in which an initial set of in-person contacts (seeds) recruit further participants from within their real-world social networks, expanding the sample iteratively through face-to-face interaction in naturalistic field settings. It is the default snowball approach in ethnographic and community fieldwork, and is particularly valuable when the target population is hidden, hard-to-reach, or lacks a sampling frame.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use field-based snowball sampling when the target population is hidden, stigmatized, or lacks a sampling frame (e.g., undocumented migrants, street-involved individuals, underground market participants), and when in-person trust-building is essential to participant recruitment. It suits qualitative and mixed-methods research in ethnography, public health, criminology, and social work. It is especially powerful when social network ties among participants are themselves analytically relevant. Do not use it when: (1) a probability sample is required for statistical inference; (2) the population is accessible through a registry or list that supports random selection; (3) the research context does not permit prolonged field presence; or (4) the referral network is so homogeneous that it systematically excludes important subgroups.
Strengths & limitations
- Grants access to hidden or hard-to-reach populations that cannot be sampled through any other practical means.
- In-person introductions leverage existing social trust, dramatically improving participation rates among wary or marginalized groups.
- Naturally maps real-world social network structures, providing relational data as a byproduct of sampling.
- Flexible and adaptive — the researcher can redirect the chain toward underrepresented subgroups as gaps emerge.
- Particularly powerful in low-resource or low-infrastructure field settings where digital outreach is infeasible.
- Non-probability design precludes statistical generalization; findings describe the sampled network, not a defined population.
- Prone to homophily bias — people refer others similar to themselves, producing a sample dominated by dense social cliques and missing peripheral or isolated members of the target population.
- Highly dependent on the quality and social reach of initial seeds; a poorly chosen seed can confine the entire sample to a single social cluster.
- Requires sustained field presence and relationship-building, making it time- and resource-intensive.
- Referral chains can stall or collapse if participants are reluctant to identify peers due to stigma, legal risk, or community norms.
Frequently asked
How is field-based snowball sampling different from respondent-driven sampling (RDS)?
Both use chain referral, but RDS is a formalized, statistically adjusted variant: it uses dual incentives (one for participating, one for recruiting), limits the number of coupons each participant can distribute, and applies mathematical weights to approximate probability-based estimates. Standard field-based snowball sampling applies none of these mechanisms and does not support population-level statistical inference in the way RDS claims to. Choose RDS when you need adjusted prevalence estimates; choose field snowball when the goal is qualitative depth or network mapping without statistical extrapolation.
How many seeds do I need?
There is no single rule, but starting with at least three to five seeds drawn from different social niches of the target population is strongly recommended. A single seed typically produces a homogeneous chain. Multiple seeds allow the researcher to compare chains and detect whether they converge on the same participants (suggesting a small, closed network) or diverge (suggesting a larger, more heterogeneous population).
When should I stop collecting referrals?
Recruitment should continue until theoretical saturation — when new field contacts are referring the same individuals already in the sample, or when new interviews produce no new themes or participant types relevant to the research question. In practice this often occurs between 20 and 50 participants, but the criterion is conceptual richness, not a fixed number.
Can I combine field-based snowball sampling with other sampling strategies?
Yes. A common approach is sequential mixed sampling: use purposive sampling to select diverse seeds, then snowball from each seed to extend reach within each social cluster. Some researchers also apply quota controls — stopping referral chains from specific subgroups once they reach a target number — to impose minimal proportional structure on an otherwise uncontrolled chain.
What ethical precautions are essential in field snowball recruitment?
Participants must never be pressured to refer peers. Referral data should be stored separately from interview data to protect anonymity. In settings involving criminalized behavior, the researcher must clarify that participation is voluntary and confidential, and must follow institutional review board protocols on limits to confidentiality. Community gatekeepers should be briefed so that recruitment does not expose participants to community-level retaliation.
Sources
- Goodman, L. A. (1961). Snowball sampling. Annals of Mathematical Statistics, 32(1), 148–170. DOI: 10.1214/aoms/1177705148 ↗
- Biernacki, P., & Waldorf, D. (1981). Snowball sampling: Problems and techniques of chain referral sampling. Sociological Methods & Research, 10(2), 141–163. DOI: 10.1177/004912418101000205 ↗
How to cite this page
ScholarGate. (2026, June 3). Field-based Snowball Sampling. ScholarGate. https://scholargate.app/en/survey-methodology/field-based-snowball-sampling
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Purposive samplingSurvey Methodology↔ compare
- Respondent-Driven SamplingSurvey Methodology↔ compare
- Snowball SamplingSurvey Methodology↔ compare