Finding participants for a research study is straightforward when your target demographic is broad.
If you need standard consumer opinions, you can simply buy a panel or run digital ads.
But when you need to interview undocumented immigrants, elite hackers, or individuals with a rare medical condition, standard recruitment methods fail entirely.
What is snowball sampling?
Snowball sampling is a non-probability sampling technique where existing study subjects recruit future subjects from among their acquaintances.
It is also known as referral sampling, chain-referral sampling, or network sampling.
Because the sample grows outward like a rolling snowball, it allows researchers to access populations that do not have a public list or database to draw from.
Instead of relying on random selection, the researcher depends on the social networks and trust of their initial subjects to find the next wave of participants.
The three types of snowball sampling
Researchers control how the sample grows by setting rules for how many people each participant is allowed to recruit.
| Type | Recruitment structure | Best used when |
|---|---|---|
| Linear snowball sampling | Each participant recruits exactly one new subject. | You want to keep the sample size small and heavily control the pace of data collection. |
| Exponential non-discriminative | Each participant recruits multiple new subjects, and all are accepted. | You need to grow the sample size rapidly and have broad inclusion criteria. |
| Exponential discriminative | Each participant recruits multiple subjects, but the researcher filters and selects only a few. | You want to reach a large pool quickly but need to balance the demographics of the final group. |
When should researchers use referral sampling?
You cannot pull a random sample if a population has no directory, registry, or census data.
Referral sampling is the practical choice when a sampling frame simply does not exist.
It is particularly useful for studying groups that are difficult to identify or hesitant to speak with outsiders.
Marginalized or hidden populations: People engaging in illicit activities or undocumented individuals often avoid public identification due to legal risk. A researcher cannot find them in a phone book, but an initial trusted contact can vouch for the researcher and bring others on board.
Highly specific niches: Members of an exclusive underground club, niche hobbyists, or professionals in highly specialized, low-headcount roles.
Rare medical conditions: Patients with a disease so rare that local hospitals cannot provide a large enough sample size. Researchers often rely on a few initial patients to refer others they know from private support networks.
Key limitations to disclose in your methodology
Because snowball sampling is not random, you cannot technically generalize your findings to the entire population.
The initial participants heavily influence the final sample, leading to significant selection bias.
If your first subject recruits only their close friends, your data will likely reflect a specific subgroup with shared traits or opinions, rather than the true diversity of the target group.
Researchers sometimes try to mitigate this by starting multiple, completely unconnected "snowballs" to capture different networks.
Expert tip: When writing your methodology section, explicitly name "selection bias" and "non-probability design" as limitations. Reviewers expect you to acknowledge that your sample is not statistically representative, so frame your findings as exploratory or descriptive rather than conclusive.
Transitioning from paper to digital surveys
Once you identify your initial participants, you need a frictionless way for them to share your survey with their network.
Paper questionnaires add unnecessary hurdles to chain-referral sampling.
Handing a participant five physical surveys to pass along to their friends rarely results in five returned responses.
Physical papers also compromise the anonymity that sensitive populations often require to participate honestly.
Moving to a digital format lets subjects simply forward a link via text message or a private chat group.
If you already have your approved methodology and questions in a document, you can use a PDF to Google Form converter to quickly digitize your static questionnaire.
This ensures your formatting stays intact while making it effortless for your snowball sample to grow organically.
FAQ
Is snowball sampling a random sampling method?
No, it is a non-probability sampling technique. Participants are not chosen randomly from a master list, meaning every member of the target population does not have an equal chance of being selected. The results depend entirely on who the initial contacts choose to recruit.
What is the difference between snowball sampling and purposive sampling?
Purposive sampling involves the researcher directly hand-picking participants based on specific traits, knowledge, or expertise. In snowball sampling, the researcher relies on the participants themselves to find and recruit the next wave of subjects. While both are non-probability methods, purposive sampling relies on the researcher's judgment, whereas snowballing relies on participant social networks.
How do you determine sample size in snowball sampling?
There is no strict mathematical formula for sample size here because the total population size is usually unknown. Researchers typically stop recruiting when they reach "data saturation" - the point where new participants stop providing new or surprising information. Alternatively, practical constraints like time, budget, or a hard deadline dictate the final sample size.
Is snowball sampling qualitative or quantitative?
It is overwhelmingly used in qualitative research, such as in-depth interviews or focus groups. While you can technically distribute a quantitative survey via referral, the lack of a random sample is a major hurdle. You cannot run standard statistical probability tests or calculate reliable margins of error on data gathered this way.
Reaching hidden communities is hard enough without getting bogged down in survey administration. By setting clear boundaries on how participants recruit each other and acknowledging the built-in biases, you can gather rich data that would otherwise remain completely out of reach. When you are ready to start collecting that data, tools like Doc2Form can turn your drafted research questions into live, shareable digital surveys in seconds, letting you focus on the people rather than the formatting.