What is systematic sampling?
Systematic sampling is a method where researchers select members of a population at a fixed interval, called the sampling interval, after a random starting point. If your interval is 10, you’d pick a random start between one and 10, then select every 10th person after that.
Say you want to learn about the people who visit a restaurant on a given day. You could station a researcher at the door and invite every 10th guest to take part. You could also use an interval of 20, 30 or whatever fits your study.
Key takeaways
- Always choose a random starting point, or your systematic sample isn’t truly random.
- Systematic sampling selects every kth person from a list or flow after a random start. Simple random sampling picks each person independently at random.
- Systematic sampling is faster and cheaper, especially for large populations or in-person intercepts.
- Patterns in your population can wreck a systematic sample if they line up with your interval.
When to use systematic sampling over simple random sampling
Use systematic sampling when you need a fast, low-cost sample and your population has no repeating pattern. Use simple random sampling when your population is small, when a hidden pattern could line up with your interval or when someone could influence who gets selected.
Both methods give every member of your population a fair shot at being chosen. They differ in how you make the picks, and that difference changes your cost, your timeline and your risk of bias.
Systematic sampling is simple to execute.
In order to perform simple random sampling, each element of the population of interest must be separately identified and selected. With systematic sampling, a sampling interval is used to select the individuals that will comprise the sample.
If researchers are working with a small population, random sampling will provide the best results.
However, if the size of the size of the sample that is required to perform the study increases, and researchers find themselves needing to create multiple samples from the population, these processes end up being extremely time-consuming and expensive.
That’s why systematic sampling is the preferred sampling technique in this scenario.
If patterns are present, avoid systematic sampling.
When there is no pattern in the data, systematic sampling is more effective than simple random sampling.
But in circumstances where the population is not random, researchers are at risk of selecting individuals to comprise their sample that possess the same characteristics, which in turn has a negative effect on data quality.
For example, if a farm that grows oranges has a sorting machine that is on the fritz, and every tenth orange that passes the sorting test is damaged, researchers are more likely to select a damaged orange to be a part of their sample if they use systematic sampling than if they were to use simple random sampling. This would result in a biased sample, and inherently poor data quality.
Use systematic sampling when there’s low risk of data manipulation.
Systematic sampling is the preferred method over simple random sampling when a study maintains a low risk of data manipulation. Data manipulation is when researchers reorder or restructure a data set, which can result in a decrease in the validity of the data.
If the risk of data manipulation is high, and the sampling interval that comes with systematic sampling has the potential to alter the data being collected, then a simple random sampling method is more appropriate and effective.
What’s the difference between systematic and simple random sampling?
The key difference is how selections connect to each other. In simple random sampling, every possible combination of people is equally likely. In systematic sampling, once you set the starting point, every other selection is locked in by the interval.
| Systematic sampling | Simple random sampling | |
|---|---|---|
| How people are selected | Every kth person after a random start | Each person chosen independently at random |
| Tools needed | A count and one random number | A random number generator and a full list |
| Needs a complete population list? | No, works with a flow of people | Yes |
| Speed and cost | Faster and cheaper | Slower and costlier at scale |
| Main risk | Patterns that line up with the interval | Logistics and cost for large populations |
| Best for | Large populations, in-person intercepts, tight budgets | Small populations, lists with possible patterns |
How to perform systematic sampling
As we outline the steps for performing systematic sampling below, let’s continue with our example from above in which researchers are interested in the population that visits a particular restaurant on a given day.
Step #1: Estimate the size of the population of interest.
The first step of performing systematic sampling is to estimate the size of the population that visits the restaurant on a given day.
The researchers would be unable to know the exact size of the population, since they cannot be 100 percent confident who will visit the restaurant during the day, but they can make some informed estimate as to how many restaurant patrons they can expect to see.
In this step, you just want to determine a rough estimate of the population size.
Step #2: Determine how many people need to be in your sample.
Next, the researchers would need to determine how large their sample needs to be.
Depending on the estimate derived from step one, the size of the sample could be ten, one hundred, or even more than that — it all depends on the desired data volume target for the research to be statistically significant.
Step #3: Divide population size by the number of people in your sample to decide how often to stop somebody.
In this step, the researchers would take the estimated population size from step one, and divide it by the number of people that need to be in the sample from step two.
The resulting number will be the sampling interval that the researchers should adhere to.
For example, if it’s estimated that 1,000 people visit the restaurant on a given day, and the researchers decide that they need a sample size of 100, then they would divide 1,000 by 100 to receive a sampling interval of ten.
This means that the researchers should stop every tenth person to ask them if they would be willing to participate in the study.
Step #4: Start by selecting a random individual.
It’s important to note that researchers should not start developing their sample by asking the first patron they see to participate in their study, as this would negatively impact data quality due to the fact that the process is therefore not random.
Instead, the researchers should begin by selecting a random individual, and from there choosing every tenth patron to enter the restaurant.
Conclusion
Systematic sampling gives you a fair, efficient way to build a sample without listing and randomly selecting every person in your population. It’s a strong choice for large audiences, in-person intercepts and tight timelines. Just check for patterns that could line up with your interval, and always start at a random point.
Whichever method you choose, the rest of your study needs to hold up too. Alchemer Survey helps you build surveys that are quick to complete, distribute them through the channels your audience uses and analyze responses in one place. That way every person you sample gets a smooth experience, and you get data you can act on
Frequently asked questions
Is systematic sampling a type of random sampling?
Yes, as long as you choose a random starting point. That random start gives every member of the population an equal chance of selection, which makes systematic sampling a probability sampling method. Starting with the first person on a list or through the door breaks that randomness.
How do you calculate a sampling interval?
Divide your population size by your desired sample size. For a population of 1,000 and a sample of 100, the interval is 10, so you’d select every 10th person after a random start between one and 10.
What is periodicity in systematic sampling?
Periodicity is a repeating pattern in your population that lines up with your sampling interval. If every 10th item shares a trait and your interval is 10, your sample could include all or none of those items. That produces a biased sample.
What are the advantages of systematic sampling?
Systematic sampling is fast, inexpensive and easy to run. It doesn’t require a random number generator for every pick, and it works with a flow of people, like visitors entering a location, even when you don’t have a complete list in advance.
What are the disadvantages of systematic sampling?
Systematic sampling is vulnerable to hidden patterns in your population, and it depends on an accurate population estimate to set the interval. It’s also easier to manipulate, since anyone who controls the order or the count can influence who gets selected.
When should you use simple random sampling instead?
Use simple random sampling when your population is small enough to list completely, when you suspect a pattern in the order of your list or when there’s a real risk someone could influence selection. It takes more effort but removes those risks.