Snowball Sampling

Snowball Sampling: Definition

So what is snowball sampling?

But what does non-probability sampling mean?

As well as snowball sampling, there are other types of non-probability sampling, such as convenience sampling; this is when researchers select a sample from a population because they are easily accessible.

Snowball Sampling in Research

We may think of snowball sampling as a kind of referral system. Researchers are prone to adopt this sampling technique when the sample size is smaller and less accessible. They ask for assistance in seeking a larger sample size by having their subjects seek out similar ones.

This sampling method can continue on and on just as a snowball continues to roll. The recruiting process can stop once the researcher has gathered enough data to analyse and draw conclusions on their hypothesis.

Snowball sampling is also an effective method for researchers when the sample of participants they require may be challenging to find.

Figure 1: Snowball sampling relies on participants referring others.

With snowball sampling, as it is non-probability-based, it must be verified before use to justify why other sampling methods with more reliability and generalisability cannot be used.

Types of Snowball Sampling

This sampling method is commonly employed in qualitative research where the group of participants is difficult to locate.

There are different types of snowball sampling techniques to consider:

Linear snowball sampling is when the research is determined by a singular, 'linear' sequence of referrals initiated by one participant and goes on. This creates a line of subjects that continues until the researcher is happy with the sample population reached.

Exponential non-discriminative snowball sampling is when the first eligible subject is approached for the study, and then conducts multiple referrals. The chain continues until enough participants are gathered.

Exponential discriminative snowball sampling differs slightly from exponential non-discriminative snowball sampling. The only difference is that this type is discriminative as the researcher screens the participants before accepting them to be involved in the research. The selection of a subject is dependent on the type of study.

Snowball Sampling Methods

The researcher calls an initial sample group to participate in the study and asks the subjects to recruit other individuals who they deem fit the criteria. If they are willing, they join the pool of participants.

The names of potential participants do not need to be shared as a matter of privacy. Individuals are continually recruited until the researcher is satisfied with the sample size.

The method used depends on the type of snowball sampling method used.

Snowball Sampling Examples

Snowball sampling as a technique can be applied to several situations requiring referrals rather than the researcher's direct recruitment of subjects.

Some individuals may be deterred from being approached by researchers because of stigma.

The snowball sampling method is used for social studies. It is a helpful research technique for individuals who may be difficult to identify.

The snowball method is used in a lot of research on specific conditions. It is useful when researching a very rare or difficult condition/situation.

Snowball Sampling Advantages and Disadvantages

There are both advantages and disadvantages to adopting snowball sampling as a technique.

One of the critical reasons researchers prefer using snowball sampling is that it enables access to the more hidden population that may not be amongst society. It also allows them to gather data without being necessarily upfront or intrusive.

Researchers save time and resources trying to piece together a sample group because other subjects are effectively responsible for finding and encouraging new participants.

Snowball sampling also follows a very ethical process. The study is approved by the Board of Institutions to ensure that it is being conducted justifiably and morally. Participants are also not forced to participate; they can accept, decline, or remain anonymous.

Because the sample is found through referral, this does mean that there is a potential for sampling bias. The whole study is highly reliant on those initial individuals participating. A specific type of person may be more likely to be referred, such as someone more sociable in that particular community. Hence, the conclusive results may not end up being representative of that population.

Because the sampling method is subjective, it would be hard to determine the sampling error.

Those involved may also not want to provide details on others participating in the study (protecting their identity or worried about ethical issues).