KANDA DATA

When conducting survey research, the number of samples observed will naturally be one of the main considerations. In survey-based studies, using samples is often a more efficient choice compared to carrying out a census on all population members. By taking a representative sample, we can observe behaviors that reflect the larger population.

By working with samples, researchers or students conducting studies can save costs, time, and effort. Moreover, as long as the samples are representative, the results will still provide a sufficiently accurate picture of the population being studied.

That said, there is always a challenge in ensuring that the sample size is truly representative. Many people wonder: What is the minimum number of samples needed in a survey study to remain representative? In this article, Kanda Data will explain how to determine the minimum representative sample size. I’ll also provide examples so readers of Kanda Data’s blog can gain a clearer understanding.

Understanding the Difference Between Probability Sampling and Non-Probability Sampling

Before learning how to determine the minimum representative sample size, there’s one important thing we need to revisit from research methodology: sampling techniques.

Sampling techniques can generally be divided into two categories: probability sampling and non-probability sampling. So, what’s the difference? Let’s break it down.

Probability sampling is defined as a sampling method where each member of the population has an equal chance of being selected as part of the sample. Examples include simple random sampling, stratified random sampling, and cluster sampling.

Meanwhile, non-probability sampling does not give every population member the same chance of being selected. Examples include purposive sampling, convenience sampling, and snowball sampling.

This distinction is crucial because the method chosen will directly influence how we determine the minimum sample size. To make this easier to understand, I’ll provide examples of how to determine minimum sample sizes for simple random sampling (probability sampling) and snowball sampling (non-probability sampling).

Minimum Sample Size in Simple Random Sampling (Probability Sampling)

In simple random sampling, the minimum sample size is usually calculated using statistical formulas. These formulas make it much easier to decide how many samples should be taken. Two of the most popular formulas are Slovin’s Formula and Cochran’s Formula.

1. Slovin’s Formula

Before applying Slovin’s or Cochran’s formulas, we must ensure the assumptions are met: namely, the population is homogeneous, and a sampling frame exists (e.g., population members are listed by name and by address).

Slovin’s formula is written as:

Example: Suppose a researcher observes a population of 1,000 people and wants to determine the minimum sample size using a 5% margin of error.

n = N / (1 + N * e^2)

n = 1000 / (1 + 1000 * 0,05^2)

n = 1000 / (1 + 1000 * 0,0025)

n = 1000 / (1 + 2,5)

n = 1000 / 3,5

n = 285,71

n = 286

So, the minimum representative sample size would be 286 people, selected randomly from the population of 1,000.

2. Cochran’s Formula

Now let’s look at Cochran’s formula, which works on a similar principle. Cochran’s formula is:

n0 = (Z^2 * p * q) / e^2

Where:

n0 = minimum sample size

Z = Z-score for the desired confidence level (e.g., 1.96 for 95%)

p = expected proportion (often 0.5 if unknown)

q = 1 – p

e = margin of error

Example: For the same case above, with a 95% confidence level (Z = 1.96), p = 0.5, q = 0.5, and e = 0.05:

Maka dapat kita hitung sebagai berikut:

n0 = (Z^2 * p * q) / e^2

n0 = (1,96^2 * 0,5 * 0,5) / 0,05^2

n0 = (3,8416 * 0,25) / 0,0025

n0 = 0,9604 / 0,0025

n0 = 384

So, using Cochran’s formula, the minimum representative sample size would be 384 respondents.

Minimum Sample Size in Snowball Sampling (Non-Probability Sampling)

Now that we’ve covered simple random sampling, let’s move on to snowball sampling, which falls under non-probability sampling. In this method, the minimum sample size cannot be determined using statistical formulas, since not all population members have an equal chance of selection. Instead, the determination is more flexible, depending on the research context.

Some common approaches include:

1. Data Saturation

Sampling stops once no new information emerges. For example, after interviewing 25 respondents, the researcher finds no new insights. At this point, 25 respondents are deemed sufficient.

2. Minimum Numbers by Convention

In qualitative research, snowball sampling often uses 20–30 respondents as a minimum, especially in exploratory studies.

3. Practical Considerations

For instance, in studies involving closed communities (e.g., people with rare diseases or niche SMEs), the sample size may be smaller (10–15 respondents) due to limited access. Example: A researcher starts with one key informant, who refers three others. Each of those three refers additional participants, and the process continues until 25 respondents are reached. At that point, no new data is emerging. Thus, the final 25 respondents are considered representative for the study’s purpose.

Conclusion

Determining the minimum sample size in survey research depends heavily on the sampling technique used. In probability sampling such as simple random sampling, formulas like Slovin or Cochran can be applied to calculate the minimum number of samples. In non-probability sampling such as snowball sampling, the sample size is instead guided by principles like data saturation, established conventions, or practical limitations in accessing respondents.

By understanding this, researchers can choose the most appropriate approach to ensure their study remains both representative and efficient.

That’s all for this article from Kanda Data. Hopefully, it provides useful insights and adds to your research knowledge. Thank you for reading, and stay tuned for the next update!