Know very little about your survey population and you just don’t know how to locate more of them? You need to conduct a business research but you just don’t have that many people in your social network to invite? Use snowball sampling. But know its advantages and disadvantages.
Snowball sampling is a form of nonprobability sampling that can be deployed as a data collection strategy when the population is hard to locate or when there is little information about the population (See Goodman, 1961; Rissell & Khavarpour, 1997). When conducting a survey study online, this snowball sampling technique can be used to ask the participants to refer the survey to their contacts or peers by forwarding the email containing the link to the survey or letting their contacts or peers know about the survey study.
Snowball sampling is a subset of a purposive sample. A purposive sample is a group of participants constructed by the researcher for a specific sampling need.
For example, as an HR researcher, you can construct in your mind a group of executives that you would like to invite to participate in a study. You want to have a good number of respondents for good statistical representation but you might not have a long list of high-level executives. However, it is very likely that the executives that you know have a larger social network of fellow high-level executives whom they can invite to participate in your study. You are essentially deploying snowball sampling technique. It is called snowball sampling because it is analogous to an accumulating snowball rolling down the hill. Your participants are basically helping you increase your participant-base. It is also called chain referral sampling or referral sampling.
One advantage in using snowball sampling in business research is that it is a convenient, quick, and relatively cost-effective way of collecting data. It is particularly helpful if researchers already have a large pool of potential participants.
One major limitation of snowball sampling is the introduction of bias. Besides the researcher constructing the sample group, the additional sampling happening in snowball sampling is respondent-driven. In the snowball sampling strategy, the sample is not randomly drawn; the sample is purposive and hence exposed to the biases of the researcher as well as the participants. It is a valid and effective technique to use in business research but extra diligence and rigor should be exercised when generalizing the findings in snowball sampling techniques or in any nonprobability sampling methods.
In : Research
Tags: research statistics
comments powered by Disqus