KANDA DATA

In conducting research, we generally take samples from a population under observation. Of course, it’s much easier if we already have data on the population size, so we can take a representative sample that reflects the population as a whole.

But what if we’re dealing with a research situation where the exact population size is unknown? Meanwhile, the research subject we want to observe is precisely that population. After some investigation, we find that the exact number is unclear—and even the presence of the population members is hard to trace.

In this article, I’ll talk about a sampling technique that’s commonly used in such situations: Snowball Sampling.

What Is Snowball Sampling?

By definition, snowball sampling is a non-probability sampling method where the researcher relies on recommendations from initial respondents to identify the next ones.

Non-probability sampling here means that not every member of the population has an equal chance of being selected as a sample. So, we begin by interviewing one respondent, then two more, and so on. Each respondent recommends others who fit the research criteria.

As the process continues, the number of respondents keeps increasing—just like a snowball rolling down a hill. That’s why this method is called snowball sampling.

Advantages and Disadvantages

Like any other sampling technique, snowball sampling has its advantages and disadvantages. Let’s start with the advantages.

• It’s very efficient for observing hidden or hard-to-reach populations where we don’t know the exact size or location.
• It’s also cost-effective, since recommendations come from respondents we’ve already interviewed.
• If the researcher already has some initial access to the community being observed, the process of gathering respondents becomes even easier.

On the flip side, this technique also has potential drawbacks:

• It may be less representative, especially if respondents can’t all be reached through recommendations from others.
• There’s a chance that respondents will only refer people similar to themselves, which may not align with the goals of the study—and this can lead to bias.

Case Study Example

To make it clearer, let me give you a case study example on how this sampling method works in practice.

Let’s say a researcher wants to study female entrepreneurs in a village. The goal is to observe how women in the village run small businesses—sewing, baking, selling vegetables, or other roles. The challenge is that the researcher doesn’t have a complete list of who these women entrepreneurs are.

Not all of them are part of a cooperative or formal group. Some might not even be active on social media or any public forum. Given this situation, snowball sampling becomes a suitable approach.

The researcher can start by contacting one well-known female figure in the village who’s active in local small business activities. After interviewing her and collecting the necessary data, the researcher then asks her to recommend 2, 3, or maybe even 5 other women who also run home-based businesses.

Then the researcher visits, say, three of the recommended women, interviews them, and again asks each of them to recommend someone they know who fits the criteria. Over time, the researcher ends up interviewing many female entrepreneurs in and around the village.

What’s important here is that the researcher makes sure all respondents meet the criteria and gives their consent to be interviewed.

Tips for Using Snowball Sampling

Here are a few practical tips to keep in mind when using snowball sampling:

1. Start with someone you can trust in the community or population you’re observing.
2. Explain your research goals clearly, so that respondents are happy to help and share the data you need.
3. Always validate the recommended respondents to make sure they actually fit the objectives of your study.

Final Thoughts

That’s a brief explanation of the snowball sampling technique, along with a real-world case study. I hope this article helps and adds value to your research experience.

Thanks for reading, and see you again in the next article on kanda data!