There are many many different ways to develop your sampling strategy. In the past I have spoken about some of these in detail, for example I mention and discuss from a holistic perspective what is sampling, and why as a concept it is extremely important. In my time as an academic I have come across many proposals, articles, workshop papers and seminars which feel incomplete due to a inappropriate sampling strategy.
In a previous related blog post, I discuss a specific but popular approach of sampling know as purposive sampling, a methodological approach which develops sampling based on specific defined criteria. Judging by the numbers of views these two posts have accumulated it is clear that there is a demand for this type of discussion in a clear and concise manner. It is to this end that today I discuss another very popular method of sampling, known as snowball sampling.
This is quite a simple approach and one that works well if built on top of a previous defined strategy as part of your research, so for example it would work well if built on top of a purposive sampling strategy, but may not be suitable as a primary sampling approach, however this would depend on your research etc. It is a form of non-probability sampling and works well based on defined criteria. In my view it maybe best viewed as a complementary strategy, helping to obtain a richer level of depth and detail, and may bring forward elements of research you did not initially consider.
In essence snowballing strategy involves asking your current respondents who have already been interviewed to identify or recommend other people they know who may fit the selection criteria (Crabtree & Miller, 1999; Groenewald, 2004). In other words, snowball sampling method is a good mechanism to obtain additional referrals from respondents to generate additional interview participants. This approach allows you to utilise the original group as informants who can help in identifying additional people to interview who qualify for inclusion and these, in turn, identify yet others – hence the term snowball sampling (Cohen et al., 2007).
It is important however that such an approach is clearly documented and for depth it maybe useful to highlight in your research the original referrer. This shows an additional level of detail and highlights a certain amount of rigour in the sampling process
Advantages of snowball sampling
The process of snowballing has several advantages, firstly as alluded to above this process is a novel way in getting access to hidden populations of your identified sample. Secondly the process can be relatively cost effective, in many cases additional sample respondents can be found in the same institution. Finally quick access to your sample can help you to complete your research quicker, giving a more efficient timeline.
Disadvantages of snowball sampling
The disadvantages of this approach include firstly oversampling. If you oversample from one particular network this may bias some of your data. There are also ethical issues, respondents may be hesitant to provide names, due to issues of anonymity. finally it is difficult to determine the sampling error and make decisions from the sample population due to the removal of random selection of samples.
Overall I hope that this blog post has helped to identify the role of snowball sampling in research data collection.
Good luck
References
Cohen, L., Manion, L., & Morrison, K. (2013). Research methods in education (7th Editio). Book, Milton Park, Abingdon, Oxon: Routledge Ltd. Retrieved from http://www.dawsonera.com.librouter.hud.ac.uk/depp/reader/protected/external/AbstractView/S9780203029053
Crabtree, B. F., & DiCicco-Bloom, B. (2006). The qualitative research interview. Medical Education, 40(4), 314–318. Journal Article.
Groenewald, T. (2004). A phenomenological research design illustrated. Journal Article.
The Academic Triangle 