The Five Survey Sampling Methods and
How to Use Excel for Sampling

Survey sampling research relies a lot on the sample quality and quantity. We’ve previously discussed how to get a good sample size for any of your target population, now it’s time to turn our attention to how to get high-quality sample subjects.

Now - what do high-quality sample subjects even mean? It’s simple; you’ll need to get a sample that truly represents your population. To do that, you’ll need to pick the right survey sampling methods for your market research needs.

In this article, we’ll be covering some of the most popular sampling methods used in market research studies. Plus, we’ll also guide you in the process of extracting samples using Microsoft Excel for random sampling methods.

Types of sampling methods

Although there are tons of sampling methods used in research contexts, most studies tend to rely on the following five sampling methods.
 

Random sampling method

Random sampling method is considered to be one of the best sampling methods for survey-based research. However, it is also known as one of the most challenging to do - probably only second to comprehensive research. At the same time, because of this, this sampling method tends to be more unbiased compared to others.

Random sampling method, in sum, is a type of sampling method where the researcher uses a random technique to select research subjects from a population. Some of these techniques are:

• Make a random number table
• Roll the dice
• Lottery

An important part of this process is that the researcher must know the members of the population being studied. Thus, it is perfect for small-scale studies such as finding out the favourite pencil brand bought by the students of High School A, for example.

In practice, it looks something like this:

The selected figure is chosen as the sample using a number generator, with each member of the population having an equal chance of being selected.
 

Stratified sampling method

Another popular sampling method in market research is the stratified sampling method. It is very similar to random sampling with one difference - the samples are grouped based on certain attributes.

Here’s an illustration of the process:

For a better example, consider the case of "investigating the work satisfaction of a corporation’s employees nationwide."

Let’s say that there are more people working in the headquarters - around 10,000 employees, while the six local branch offices, only have 1,000 each. So, the population is around 16,000 employees, and the researcher decides to take 1,600 employees as the sample size.

If the researcher takes 1,500 people from the headquarters and just 100 from all of the local branches as the sample, the survey’s response would not represent the entire corporation - just the headquarters employees.

This is when a stratified sampling method will be useful. For a sample size of 1,600 people, the researcher has to include at least 600 people from local branch offices, with 100 taken from each.

Since sampling is performed for each layer with similar attributes, the bias of the results is small within each layer (within-group), but the variation between layers (each group) is large.

Remember, that as with the random sampling method, when using this method, you need to know the members of your target population.
 

Multi-stage sampling method

Multi-stage sampling method is best used in large-scale research. It simplifies the process of a typical random sampling by taking a sample of the larger groups and scaling it down into smaller groups through subsequent random sampling at each stage.

It typically looks like the following:

One example of this sampling method can be seen in a policy awareness survey conducted in each region of the country. Conducting a national awareness survey requires a lot of human, time, and economic costs.

But costs can be reduced by performing multi-stage sampling according to the following procedure:

In the end, the researcher would only need to conduct a survey on the 300 households.

The downside of this method is that the sample might be more prone to bias due to the small sample size in comparison to the population.
 

Cluster sampling method

Cluster sampling method is actually quite similar to the multi-stage sampling method. There’s one caveat though; it is nowhere as complex as the latter.

In cluster sampling, researchers would collect groups with similar characteristics to the population being studied. It’s a bit different from multi-stage sampling, where the intent is to get the smallest possible group of people who can represent the population.

Imagine that you’re trying to determine the most popular instant noodle brand amongst first-year high school students in a certain area where 25 high schools exist. If your sample size is five, you’ll just need to use random sampling to select five high schools in the area from the existing population.

 

Systematic sampling method

Systematic sampling is a type of sampling method where the research participants are selected based on a fixed number interval.

Here’s an example. You’re trying to find out about how people usually choose the movies they’re watching at a local cinema chain. So, you decide to survey every 10th person who is queueing at the ticket box from the opening to closing hours.

However, if the elements of the population are arranged in a certain order, such as alphabetically or by grade, the sample taken from the population may be biased.

How to Use Excel for Sampling

All of the sampling methods discussed above involve some random sampling techniques as part of the process. While it is possible to generate a random sample using written calculation, it isn’t the most efficient way to do so.

The best way to generate a random sample would be through using Microsoft Excel. The process is actually quite simple:

Summary

Aside from sample size, another factor that you shouldn’t overlook while gathering your sample is choosing the right sampling method. A good sampling method reduces the risk of bias and in the process, increases the validity and reliability of your research results.