In order for these surveys to yield meaningful results, finding the right sample size is key. And although many people don't like math, in business, understanding and collecting data is reliant on this subject.
In this post, we'll help you learn what survey sample size you might need to have a statistically significant result from your surveys.
What are survey sample sizes?
A survey sample size refers to the number of participants you have in a study because it's not usually possible to gather responses from an entire target group. It's important to have the right sample size so your findings will hold statistical significance, meaning you can confidently draw conclusions from the study about a certain population.
These are called sample sizes because they are only a small representation of a larger group of people who you want to know more about. You can choose who to include in your sample or you can conduct a random sample where respondents are chosen by chance.
Survey sample sizes that are too small may include a disproportionate amount of outliers. When you try to gather responses from too many people, although the results will be more accurate, the study will become too complex, expensive, and time-consuming.
What is a good sample size?
While there isn’t a definitive number for your sample size, experts say it should be at least 100 if you want a shot at meaningful results. So if your population is 100 or less, you’ll need to survey all of them.
Ideally, your maximum sample size should be no more than 10% of the population, but not exceed 1,000. For example, if your population size is 30,000, 10% would be 3,000 — exceeding the 1,000 threshold.
For a more accurate way to determine your sample size, you can utilize a formula. Let me warn you: the sample size formula can be a little overwhelming. But we'll work through it together. Let's review it below.
That being said, how many people should you include in your sample?
Sample Size Formula
The formula above is known as Slovin’s Formula. I know that looks daunting….The good news is that you won't have to do this math yourself. There are plenty of sample size calculators, including one from HubSpot in our A/B Testing Kit. However, it's a good idea to be familiar with this formula and what it means.
The elements of this formula are:
- z = z-score. The number of standard deviations a given proportion is away from the mean.
- p = standard of deviation. The confidence level you can have that the population would select an answer within a certain range.
- e = margin of error (percentage in decimal form). The percentage that tells you how much you can expect your survey results to reflect the views of the overall population.
- N = population size. The total number of people in the group you're trying to study.
To find your z-score, use the table below.
Essentially, your sample size is decided based on the population, how confident you want to be in your results, and the margin of error you'd like to allow for. Let's look at an example.
Below, I used a sample size calculator and I put in that my population size is 10,000, and I'd like to be 95% confident in my results within a 5% margin of error. The calculator then used the formula above and told me my sample size needs to be at least 370 people.
Now that we know the formula (and how to find a calculator to do it for you), let's review some steps that you should take to ensure your results are statistically significant.
How to Find Sample Sizes
- Know your population.
- Have a high confidence level with a low margin of error.
- Plan for a response rate of about 20-30%.
- Estimate the standard deviation.
- Troubleshoot sample size results.
1. Know your population.
The population size in your formula should be the number of the entire group of people you want to understand. If you want to know how many people in the U.S. like your product, you obviously can't survey over 300 million people. However, it's important to know the total number of people that you want to understand — whether it's people in a country, from a certain company, or in a certain profession.
2. Have a high confidence level with a low margin of error.
When you have your results, you want to be confident that most of your population would agree with the results, with a low margin of error. Usually, the margin of error is around 5% or lower and the confidence level is around 95% or higher. This would mean that your results can give or take 5% on the top or bottom and that you're confident you would get the same results 95% of the time.
3. Plan for a response rate of about 20-30%.
Once you've put your information into a sample size calculator and figured out how large your sample size needs to be, it's important to plan for your response rate to be around 20-30%. Of course, not everyone you send a survey to will respond, so you want to account for that. In fact, an even more conservative approach would be to account for a 10-15% response rate, meaning you should send your survey to many more people than the required sample size to ensure you receive that many responses.
4. Estimate the standard deviation.
A low standard deviation means that the results will be clustered around the same mean number while a high deviation indicates that the results are spread across a much greater range. Since this step requires you to estimate how the responses to your survey will vary, you’ll want to pick a standard deviation of .5 to ensure your sample size is large enough.
5. Troubleshoot your sample size results.
You’re almost at the finish line but may hit some snags. If your sample size is too large, you can make adjustments by increasing your margin of error or decreasing your confidence level.
Additionally, if it is too small, sample more of your population.
Optimal Sample Sizes are Key for Meaningful Results
We know that survey sample sizes can be overwhelming, but it's important for you to understand how confident you can be in the results of your surveys, and within what margin of error. Luckily, there are plenty of survey templates and sample size calculators available to help you get started.
Editor's note: This article was originally published in December 2021 and has been updated for comprehensiveness.