Snowball sampling is a unique non-probability sampling method researchers can utilize when the target population members are hard-to-reach or difficult to identify. Also known as chain sampling, chain-referral sampling, or chain-referral, this snowball sampling method starts with a small set of initial research participants.
The researcher may then ask these participants to refer potential participants or potential subjects who meet the study’s criteria. This method allows researchers to conduct research in challenging contexts where traditional sampling approaches may not work, as one subject leads to the next, gradually expanding the participant pool.
This technique is especially valuable in research methods where the population is difficult to access, as the snowball sampling method relies on social networks to locate individuals who may otherwise be missed in more structured sampling methods.
Example: Snowball sampling
You are conducting a study on the challenges faced by entrepreneurs in a new city. Since you are new to the city, you need help locating entrepreneurs independently. You begin by interviewing entrepreneurs within your network.
Criteria for selection:
- The entrepreneur must have started their business at least three years ago.
- The business must be currently operational.
- The business must be located within the city limits.
- The entrepreneur must have experienced significant business challenges, such as securing funding or scaling the business.
When to Use Snowball Sampling
Snowball sampling is commonly used in qualitative research, especially when studying populations that are difficult to reach. These populations may include:
- Groups that are small in comparison to the general population
- Populations spread across wide geographic areas
- Groups with a social stigma or specific shared characteristic of interest
In these situations, accessing population members is challenging for outsiders because there is no available sampling frame.
Snowball sampling is frequently used in fields like public health (e.g., drug users), public policy (e.g., undocumented immigrants), or niche areas (e.g., buskers). It is also effective for studying sensitive topics that individuals might not want to discuss publicly due to perceived risks associated with self-disclosure. This method allows researchers to access these populations while addressing ethical concerns, such as maintaining privacy and ensuring confidentiality.
Types of Snowball Sampling
There are three main variations of the snowball sampling approach:
- Linear snowball sampling
- Exponential non-discriminative snowball sampling
- Exponential discriminative snowball sampling
Linear Snowball Sampling
In this basic form, each participant is asked to refer one additional person to the study. The sample size grows linearly.
Example: Linear snowball sampling
Researchers studying the challenges homeless youth face may find it difficult to reach this population through traditional recruitment methods. Homeless youth can be a highly mobile and marginalized group, often avoiding contact with authorities or social services. Using snowball sampling, the researchers could start by connecting with a local homeless youth shelter and interviewing a few young people there. These initial participants could then provide referrals to other homeless youth they know, allowing the researchers to gradually build a sample and gain deeper insights into the lived experiences of this vulnerable population.
Exponential Non-Discriminative Snowball Sampling
Participants are asked to refer multiple people, leading to exponential sample growth. All referrals are included, regardless of their characteristics.
Example: Exponential non-discriminative snowball sampling
You are studying the informal support systems utilized by new mothers. You begin by interviewing a few new mothers in your local community. At the end of each interview, you ask the participants to provide the names and contact information of other new mothers they know who may be interested in participating in the study. This exponential, non-discriminative approach allows you to rapidly expand your sample, as each new participant is asked to refer multiple acquaintances. Over time, you can build a diverse network of 40-50 new mothers who share their experiences and coping strategies.
Exponential Discriminative Snowball Sampling
Like the exponential non-discriminative approach, the researcher sets specific criteria for the types of referrals they will accept, such as only including participants from a specific demographic group or with particular experiences.
The choice of snowball sampling technique will depend on the research goals, the expected size and diversity of the target population, and the resources available to the researcher.
Example: Exponential discriminative snowball sampling
You are researching mentorship’s role in supporting immigrant entrepreneurs’ success. Beginning with a small number of immigrant business owners you have identified through local community organizations, you conduct interviews and ask each participant to refer other immigrant entrepreneurs they know who have had a meaningful mentoring relationship. However, you specify that you are only interested in speaking with those who have been running their business for at least 3 years, as you want to focus on the long-term impact of mentorship. This discriminative snowball sampling allows you to build a sample of 15-20 experienced immigrant entrepreneurs with diverse mentoring experiences.
Advantages and Disadvantages of Snowball Sampling Research Methods
Snowball sampling is a non-probability sampling technique used in research. In this technique, existing study subjects recruit future subjects from among their acquaintances.
Here is a table outlining the advantages and disadvantages of snowball sampling:
| Advantages | Disadvantages |
| Cost-effective and efficient, especially when studying hard-to-reach populations | Sampling bias, as the initial subjects tend to nominate people they know well, leading to a sample that is not representative of the entire population |
| Helps to locate hidden populations, such as drug users or sex workers | Lack of control over the sampling method, making it difficult to determine the sampling error and make statistical inferences from the sample to the population |
| Builds trust and rapport with the target population, as referrals are made by acquaintances or peers | Oversampling of a particular network of peers, leading to a lack of diversity in the sample |
| Requires little planning and few personnel compared to other sampling methods | Difficulty in verifying the eligibility of potential respondents, as they are not directly recruited by the researcher |
| The snowball effect can quickly generate a large sample size | Ethical concerns, particularly when studying sensitive topics or vulnerable populations, as the referral process may compromise privacy and confidentiality |