If you're interested in researching a specific group of people, like restaurant managers or crime victims, you need to recruit participants for your study. Calling or mailing letters to random people in the phone book would take ages, since most people you contact won't meet your criteria. The solution is snowball sampling, a practice in which a small group of initial participants help researchers find more participants by accessing their social networks. Snowball sampling is convenient, but has significant limitations.
The biggest advantage of snowball sampling is that it helps researchers find more participants for studies than would be possible by other methods. A researcher hoping to survey business executives, for example, might struggle to convince busy CEOs to take the time to fill out a survey. If the researcher happens to know a particular CEO, however, that executive could contact other people in her social network who hold the same title. Since this technique uses existing social relationships, researchers have a better chance to recruit new participants.
Reaching Hard-to-Find Populations
Snowball sampling is particularly useful for finding participants who meet unusual criteria and are typically hard to locate. For example, a researcher interested in surveying survivors of sexual assault might have a very difficult time finding participants, since many survivors never report their assaults to the police and records are often sealed from the public. However, some sexual assault survivors attend support groups and a researcher in contact with one survivor might be able to use that participant's social network to contact more survivors. This allows a larger sample size than would be available otherwise.
One problem with snowball sampling is that participants' social networks aren't random. As a result, the technique can result in biased samples. In the CEO study, for example, if the first participant recruited happens to work in manufacturing, most of her CEO contacts are likely to also come from the manufacturing sector. Thus, the study could exclude or under-represent CEOs from different industries. Similarly, a sexual assault survivor's social network might be limited to a single geographic area, making it impossible for the researcher to generalize study results to the entire nation.
Research Ethics Issues
Snowball sampling can present significant ethical problems, especially if the research topic is sensitive or personal. For example, many sexual assault survivors are reluctant to talk about their experiences and avoid identifying themselves as survivors. A researcher using snowball sampling can keep participants anonymous in published papers but can't keep participants anonymous from each other since participants have to recruit each other. Some study participants may object to using their social networks in this way, and some potential participants may dislike being contacted for participation.
- mikdam/iStock/Getty Images