Purposive and Snowball Sampling Explained

When a research question targets a specific kind of person, random selection stops making sense. You don't want just anyone. You want department chairs, or nurses who've worked night shifts, or people living with a rare condition. Two non-probability methods are built for exactly this. Purposive sampling selects participants because they fit the research question. Snowball sampling grows a sample through referrals, reaching people who would otherwise be hard or impossible to find. They're often used together, and both rest on the researcher's judgment rather than chance.

This guide explains each method, its main variants, where it fits, and the limits you need to defend in your methodology section. Purposive and snowball sampling are two of the four non-probability methods. For how they sit alongside the others, see our overview of non-probability sampling. For the broader framework, see the research methodology guide.

What Is Purposive Sampling?

Purposive sampling, sometimes called judgmental or selective sampling, deliberately recruits participants who have characteristics relevant to the research question. The logic is the opposite of random selection. Instead of giving everyone an equal chance, the researcher intentionally seeks out the people who can speak to the topic with the most depth. It's the standard approach in qualitative research, where information richness matters more than statistical representativeness.

The method stands or falls on the inclusion criteria. Because the researcher decides who fits, those criteria need to be defined clearly and applied consistently. A study of first-generation graduate students has to specify what counts as first-generation, then apply that definition to every recruit. Vague or shifting criteria are where purposive sampling loses credibility with reviewers.

Common Types of Purposive Sampling

Purposive sampling isn't a single technique. Several variants serve different research goals.

• Maximum variation sampling. Selects participants across a wide range of a characteristic to capture diverse perspectives. A study of remote work might deliberately recruit people in very different roles, industries, and household setups.
• Homogeneous sampling. Selects participants who share a defining trait, to study that group in depth. A study might focus only on first-year international PhD students in engineering.
• Typical case sampling. Selects participants who represent what's normal or average for the group, useful for describing a common experience.
• Extreme or deviant case sampling. Selects unusual or outlier cases that illuminate the boundaries of a phenomenon, such as unusually successful or unusually struggling programs.
• Expert sampling. Selects participants with specialized knowledge, common when the research depends on professional or technical expertise.

Example of Purposive Sampling

What Is Snowball Sampling?

Snowball sampling recruits initial participants, then asks them to refer others who qualify. Those referrals refer more people, and the sample grows like a snowball rolling downhill. It's the go-to method for hard-to-reach or hidden populations: people whose membership isn't listed anywhere and who may be reluctant to come forward to a stranger. Referral from someone they trust lowers that barrier.

The method trades reach for representativeness. It can access populations no sampling frame could capture, but the sample is shaped by the social networks it travels through. People tend to refer others like themselves, so the sample can cluster around one part of the population and miss the rest. That tendency is the method's defining weakness, and it has a name: network bias.

When Snowball Sampling Fits

• Hidden populations. Groups with no public listing, such as undocumented workers or people who use a stigmatized service, can often only be reached through trusted referral.
• Rare characteristics. When the trait under study is uncommon, referral chains find qualifying participants faster than broad recruitment.
• Tight-knit communities. Specialized professional networks or close communities where members know one another lend themselves to referral-based recruitment.
• Trust-dependent topics. Sensitive subjects where a personal introduction makes participants more willing to take part.

Example of Snowball Sampling

Purposive vs Snowball Sampling: Side by Side

The two methods solve different problems. Purposive sampling answers "who fits the question?" Snowball sampling answers "how do I reach people I can't find?" The table compares them across the features that matter for a methodology section.

In practice the two often work as a pair. A researcher purposively selects the first participants to make sure they fit the criteria, then lets snowball referrals carry the sample into parts of the population she couldn't reach directly. The convenience method is a third non-probability option for when ease of access drives recruitment. See our guide on convenience sampling for that comparison.

Limitations You Have to Defend

Both methods are legitimate, and both draw scrutiny. Reviewers accept them when the write-up names the trade-offs. Three limitations need explicit attention.

Limited generalizability

Neither method supports formal statistical inference to the population, because selection isn't random. What they support is analytic generalization: the argument that findings may transfer to similar contexts. State that boundary plainly so no reader assumes the results project to the whole population.

Researcher bias in purposive sampling

Because the researcher chooses who fits, personal assumptions can shape the sample. Selecting only the participants who confirm an expected pattern is a real risk. Pre-set inclusion criteria, applied consistently and reported in full, are the defense. They show the selection followed a rule rather than a hunch. For the wider set of biases that threaten validity, see our research bias guide.

Network bias in snowball sampling

Referral chains follow social ties, so the sample inherits the shape of those networks. If the first participants share a background, their referrals often do too, and whole segments of the population may never be reached. Seeding the sample from several unconnected starting points reduces the problem. Naming the likely clustering in the limitations section addresses what remains.

How to Report Purposive and Snowball Sampling

A clear methodology section makes either method defensible. Work through these steps when you write it up.

1. Name the method. State whether you used purposive sampling, snowball sampling, or both. If you combined them, explain how the two stages fit together.
2. Justify the choice. Tie the method to the research question. Explain why information-rich cases or referral-based recruitment suited your topic better than a random draw.
3. State the inclusion criteria. For purposive sampling, define exactly who qualified and how you applied the definition. This is the heart of the method's credibility.
4. Describe the referral process. For snowball sampling, report how many seeds you started with, how referral worked, and how many waves the chains ran.
5. Report the sample size and its basis. Give the final number and the logic behind it, usually saturation for qualitative work.
6. Acknowledge the limits. Name researcher bias for purposive sampling, network bias for snowball sampling, and the boundary on generalization, before a reviewer raises them.

For setting a defensible sample size, including how saturation works in qualitative studies, see our guide on how to calculate sample size for your study.

Frequently Asked Questions

Editing Support for Your Methodology Section

Your sampling method is one of the first things a committee or reviewer reads, and purposive or snowball recruitment described loosely is a fast route to a revision request. The methods themselves are rarely the problem. The problem is writing about them in a way that names the inclusion criteria, the referral process, and the bias risks precisely, without either overclaiming or burying the study in caveats.

Editor World provides dissertation editing and academic editing services for researchers writing up methodology sections, theses, and journal submissions. Every editor is a native English speaker from the United States, the United Kingdom, or Canada, with an advanced degree in their field, and every document is reviewed by a real person, never by AI. You can choose your own editor from the Editor World roster, or request a free sample edit of your first 300 words before committing. Pricing is transparent through an instant price calculator that shows your exact cost upfront. A certificate of editing confirming human-only native English editing is available as an optional add-on for journal submissions where AI use must be disclosed.

For related methodology topics, see our overview of non-probability sampling, our guide on convenience sampling, the research methodology guide, and population vs sample in research.

This article was reviewed by the Editor World editorial team. Editor World, founded in 2010 by Patti Fisher, PhD, provides professional editing and proofreading services for graduate students, academics, and researchers worldwide. Recommended by the Boston University Economics Department. BBB A+ accredited since 2010 with 5.0/5 Google Reviews and 5.0/5 Facebook Reviews. More than 100 million words edited for over 8,000 clients in 65+ countries. Page last reviewed June 2026.