In the non-probability sampling method, we use Snowball sampling to collect samples from the community for research work. As in non-probability sampling, we shortlist our samples that fall into specific criteria, so it becomes very difficult to collect samples with specific traits.
In snowball sampling or network sampling, the sample selection process starts with the research participants, whom we call representatives. Then they find more samples in their surroundings. This process continues like a chain process from one person to a group of samples.
Researchers or
dissertation writers can use snowball sampling and collect maximum samples in a short span of time. We can use the example of snowball for better understanding. We have seen that when a snowball begins rolling, it’s very small and becomes bigger as it keeps rolling. This is how snowball sampling works and gathers maximum samples from the community.
In this article, you will learn about further types of snowball sampling with examples.
Why is snowball sampling significant?
Sometimes, when the research objective is related to illegal activities like drug addiction, cheating on exams, prostitution or snatching, people do not want to become part of the research. So, data collection becomes a very difficult process. In such cases, snowball sampling representatives play a vital role in finding people in their surroundings and recruiting them to study their behaviour. Advantages
- It makes data collection possible from those areas where direct reach is impossible due to less number of research participants
- You can also discover the characteristics of communities that are away from your access
- It is a cheap and easy method of biased sampling
- It is a flexible method, as only those become samples who want to participate
Disadvantages
- It is a biased sample collection method, so there are chances of errors in results about an entire population
- Researcher is dependent on the participants, and the circle of his research scale is dependent on the reach of the participants
- Research also relies on referrals and the researcher does not know them personally. So, referrals may fail to refer more people, or people feel hesitation to trust in referrals.
Types of snowball sampling
The sample collection process starts with the initial members of the research. We name these initial samples “seeds”. These seeds recruit more samples from their community, and this process continues in the form of waves. Samples of wave 1 find more samples and make wave 2. The sample collection process keeps growing wave by wave. There are three types of snowball sampling:
- Linear sampling
- Exponential non-discriminative sampling
- Exponential discriminative sampling
You can choose a suitable type of sampling according to your research objective.
Linear sampling
In this type of sampling, the
researcher becomes the only representative and recruits a sample directly. Then this sample finds another reference, and the process continues until the researcher collects enough samples for the research objective.
We can get desired results using linear sampling by applying restrictions about who is included or excluded.
Example:
Let’s imagine your research objective is to find out why students do cheating. You look around in your community and find a student involved in cheating on an exam. You interview him, and at the end, you request him to find other students who do cheating. Next time, he meets you with 2 other students. Conduct a collective interview of all three and at the end, ask them how they know each other.
Exponential non-discriminative sampling
In exponential non-discriminative sampling, the first 2 participants find several participants from their class and friend circle. In simple words, researchers find these two participants, and now these two have found many others. The number of samples grows exponentially in this type of sampling.
Example:
Come back to the previous example of two students. The researcher asked them to find more students around them. They introduce you to several other students, and you interview them too. At the end of the interview, you request them to look around in the other classes of their school and find more students.
Exponential discriminative sampling
In this type, the researcher interviews all the referrals of the last wave but, this time, shortlists all samples according to his criteria. Only include those referrals into samples that fall in the criteria and exclude others. It is the major difference between
discriminative and non-discriminative types. The purpose of this type is the screening of all the referrals according to the criteria to extract the desired results.
Example:
Last time you asked questions from 2nd wave and they have found several other students by their reference. Previously, you didn’t mention any criteria or limitations like the specific subject. Your objective of the research is to understand why students do cheating in mathematics or science exams. Hence, only shortlist those students for sampling who do cheating in mathematics or science exams and refuse the remaining students. After inquiring about them, you ask to find more students, and this process continues until you collect the maximum number of samples that fall within your criteria.
Final Thoughts
This article aims to clarify your concepts about snowball sampling. We explained the definition with the example of a rolling snowball and later highlighted the unique purpose of using a snowball for sampling. Researchers use this method to collect samples for their research on unacceptable topics in which people hesitate to participate. Further, we tried to differentiate it from other methods by explaining the advantages and disadvantages. Later, you studied the three types of snowball sampling, which you can use to collect maximum numbers of samples and get desired results.
