Sampling in qualitative research is typically purposeful (also called purposive), meaning that rather than aiming for statistical representativeness, researchers intentionally select participants who can provide rich, relevant and meaningful information about the phenomenon being studied. Qualitative sampling prioritises depth, rather than breadth, and the aim is to find participants who have specific experiences, characteristics or knowledge that will help answer the research question. Sampling decisions are therefore informed by the study design, research question and, in some approaches, emerging findings.
This page outlines seven commonly used purposeful sampling methods, before concluding with a brief discussion of sample size.
Criterion sampling is a purposeful sampling method that involves selecting all participants who meet a predetermined set of criteria relevant to the event, condition, role, time frame or experience being studied. For example, a study might include all patients who are readmitted to an intensive care unit within three weeks for the same complaint.
Criterion sampling is particularly useful when the aim is to explore a clearly defined experience or process in depth, as participants share key characteristics that are directly relevant to the research question.
Extreme or deviant case sampling is a purposeful sampling method that involves selecting participants who are unusual or outliers in relation to the phenomenon being studied. These cases may differ markedly from what is considered typical or average. For example, a study exploring recovery from injury might include individuals who recover exceptionally quickly or those whose recovery takes an unusually long time.
Extreme or deviant case sampling is particularly useful when the aim is to gain new insights into a phenomenon, as examining unusual cases can highlight factors or processes that may not be as apparent in more typical cases.
Maximum variation sampling is a purposeful sampling method that involves selecting participants who differ across key characteristics relevant to the phenomenon being studied. The aim is to capture a wide range of perspectives, experiences or contexts within the sample. For example, a study exploring recovery from injury might include people who recover quickly, those who take longer to recover, and individuals from different age groups or backgrounds.
Maximum variation sampling is particularly useful when the aim is to understand how a phenomenon is experienced in different circumstances, including both similarities and differences across participants.
Homogenous group sampling is a purposeful sampling method that involves selecting participants who share similar characteristics or experiences relevant to the phenomenon being studied. By minimising variation within the sample, the researcher can explore a particular experience or process in greater depth and detail. For example, a study exploring the experiences of returning to work after an injury might include participants who are all within the same age group, occupation, or stage of recovery.
Homogenous group sampling is particularly useful when the aim is to gain a detailed understanding of a specific experience within a clearly defined group. It is also commonly used in focus group research, where participants with similar backgrounds or experiences may feel more comfortable discussing shared issues.
Snowball sampling is a purposeful sampling method in which existing participants are asked to identify or recruit other people who may be willing to take part in the research. For example, a study exploring the perspectives of people experiencing homelessness might begin with a small number of participants recruited through a support service, then these participants could suggest others with similar experiences who might be interested in participating.
Snowball sampling is particularly useful when the population of interest may be difficult to access directly, such as groups that are socially marginalised, dispersed, or not easily identifiable. However, because participants are recruited through existing networks, the resulting sample might share similar characteristics.
Theoretical sampling is a purposeful sampling method used in grounded theory. Participants are selected based on their potential to contribute to the development and refinement of the emerging theory, and as data collection and analysis progresses the researcher seeks additional participants who can help explore, clarify or challenge the emerging ideas. In this way, decisions about who to sample next are guided by the developing analysis. For example, a grounded theory study exploring recovery after injury might initially include a small group of participants. Then as patterns begin to emerge, the researcher may recruit additional participants with different recovery experiences to better understand the conditions under which the patterns occur.
Theoretical sampling is particularly useful when the aim is to develop a theory that explains a process or phenomenon.
Triangulated sampling is a purposeful sampling method in which multiple sampling strategies are combined within a single study. This allows the researcher to capture different perspectives, experiences, or contexts and can enhance the depth and breadth of the data. For example, a study on mental health might begin with participants selected for shared characteristics (homogenous group sampling) and then expand the sample through participants’ networks (snowball sampling).
Triangulated sampling is particularly useful when the aim is to gather rich, diverse data and when a single sampling method might not fully capture the phenomenon of interest. By combining approaches, researchers can access multiple participant networks and increase the variation and credibility of the findings.
If you would like to consolidate your understanding of the sampling methods outlined above, you may find the following activity helpful:
In qualitative research, the sample size does not need to be large. Instead, it should be sufficient to support a detailed and meaningful analysis of the phenomenon under study. One common guide is data saturation, which occurs when collecting additional data no longer provides new insights, perspectives, or themes.
Sample size can also be influenced by the chosen sampling strategy (for example, homogenous versus maximum variation sampling), the complexity of the research question, and practical considerations such as time, resources, and access to participants. The key is to include enough participants to thoroughly explore the phenomenon while remaining feasible and ethical.