Things You Need To Know About Survey Sample Size

It always feels great when I sit down to pen down my thoughts, ideas, and knowledge with you. I have been writing about surveys for over a year, but my learning about surveys has not stopped. I enjoy learning about surveys as much as I enjoy writing about it.

When there are umpteen resources online to learn about survey sample size, why should someone read my article on survey sample size? Am I really justifying the opportunity given to write about surveys? Do I know how to write about surveys? — a few questions I ask myself.

Learning is a continuous process. I spend daily time reading about surveys as much as I spend writing about them. I read, I digest, I think, and I write. The overall idea is to provide value to people who spend time reading my article.

I am sitting with my laptop again to write on this new topic - survey sample size.

I will put my best foot forward to write on this topic to fulfill the below bullets.

• glued to content
• read till end
• leave with valuable takeaways

Let’s start.

First things first. What is a survey?

Surveys are used to receive opinions or feedback from people on a particular subject.

1. What is a survey sample size?

You may want to ask for opinions from 50 people or from 50 thousand people.

Asking for opinions from 50 people is easier than asking it from 50 thousand people.

So, what would you do if you needed the opinion of 50 thousand people?

Here is where the concept of Survey Sample Size creeps in.

The survey sample size is the size of a smaller number of people from whom, when opinions are collected, can more accurately represent the opinion of the entire targeted population of your survey.

Having discussed the sample size now is the time to know what looks like a good sample size.

2. Choosing a good sample size

• Minimum sample size: It is a general thought that the sample size should be a minimum of 100. This means that if your total survey population is less than 100, you must survey the whole population.
• Maximum sample size: Another general mode of thought is that your sample size should be 10% of your total population and, at the same time, not exceed 1000. This is because the higher your sample size, the higher the time and cost involved for that survey. For example, if your survey population is 50000, 10% of that is 5000, which exceeds the 1000 threshold.

However, there are no hard and fast rules for determining the sample size.

There is a mathematical formula to calculate the survey sample size.

I want to tell you beforehand that the sample size calculation is not simple math.

However, I have tried to explain that to you as simply as possible.

Read on and get a grasp on the calculation behind sample size.

3. How to calculate survey sample size?

As mentioned earlier, the calculation of the survey sample size involves a mathematical formula.

First, let me try to explain the terminologies involved in this formula, below.

• Population Size (N) refers to the total number of people in your population. For example, if you are surveying an organization, population size refers to the total number of employees there.

• Confidence level & z-score (z) mean how confident you are that the responses you collect are accurate. Confidence level is usually expressed in percentage.

  You need to turn your confidence level into a z-score (z) by referring to the table below. For example, the 95% confidence level has a z-score of 1.96.

Desired Confidence Level	z-score
80%	1.28
85%	1.44
90%	1.65
95%	1.96
99%	2.58

• Margin of error (ε) represents how much error you are willing to allow in your survey.

  Errors are unavoidable, and that’s why the concept of error margin.

  It is usually represented in percentages.

• Standard deviation (p^) refers to how confident you are that a population selects an answer within a given range.

  A lower standard deviation means you don’t expect your respondents to choose extreme answers.

  A higher standard deviation means you expect respondents to choose answers from the end of the spectrum.

  Standard deviation is given as a number between 0 and 1. If you are not sure about what to choose, go with 0.5, which is the exact midpoint.

With all the inputs ready, the sample size can be calculated with the formula as per the screenshot below.

Do also remember to check out the sample size calculator below.

→ BlockSurvey sample size calculator ←

This calculator is easier and faster.

I have worked out an example in the next section for sample size calculation.

4. A worked example

When you know the total population size, you can very well use the BlockSurvey Sample Size calculator mentioned earlier to find your sample size.

When your target population is unlimited or unknown, you can use the mathematical formula (for unlimited population) mentioned earlier in the screenshot of the previous section.

Let’s assume you have

• 95% confidence level (the corresponding z-score is 1.96)
• 0.5 standard deviation (midpoint chosen)
• 5% margin of error

So, the calculation goes like this.

((z-score)2 x standard-deviation(1-standard deviation)) / (margin of error)2

((1.96)2 x .5(.5)) / (.05)2

(3.8416 x .25) / .0025

.9604 / .0025

384.16 ~ 385

385 respondents are needed.

Hurray! You have just determined your sample size.

Just because 385 is a surveyable population, we are not going to the next formula and stop here.

5. Usecase-based Sample Size

Below is a table that shows the accuracy of the results.

	Increased value	Decreased value
Population size	Less accurate	More accurate
Sample size	More accurate	Less accurate
Confidence level	More accurate	Less accurate
Margin of error	Less accurate	More accurate

The larger the sample size, the larger the accuracy of the results.

We call sample sizes with greater accuracy a statistically significant sample size.

Now we’ll explore use cases requiring a statistically significant sample size and those not.

• Employee Surveys: Employee surveys in organizations are gathered to receive feedback on work culture. It is important to get a holistic view of all employees.

  In such a scenario, statistically significant sample sizes are required.

• Customer Surveys: Customer surveys are usually received on a case-to-case basis rather than in a single bulk. In this case, every customer's feedback is valuable in itself.

  In this scenario, statistically significant sample sizes are not required.

• Healthcare Surveys: Healthcare surveys are of 2 types it can be a drug discovery survey or a patient feedback survey.

  In the case of drug discovery, we need a holistic view and so statistically significant sample sizes are required.

  In the case of patient feedback, significant sample sizes are not very important**.**

• Market Surveys: Market research surveys and poll surveys need a holistic view to make decisions.

  Just like Employee Surveys, Market surveys also need statistically significant sample sizes to get an overall opinion of the entire target population.

• Education Surveys: If the purpose of your education survey is to bring changes to the school or college system, then a statistically significant sample size is important.

  If the purpose is just to know what the student thinks about the school and teachers, then a statistically significant sample size is not very important.

• Casual Surveys: If your intention is just to try out BlockSurvey or survey your friends and relatives for fun, you can ignore the significant sample size and have fun with the responses from your friends and relatives.

Now, let us move on to learn about the types of sampling.

Type of Sampling

Having learned about sample size, now is the time to learn how to select those samples.

I am discussing here 3 methods that I came across. They are random, stratified and convenience.

• Random Sampling: When you select your samples randomly from the whole population, we call it Random Sampling. Random sampling is best for the large target population.

  The responses obtained through random sampling are a very good representative of the whole population. Also, selecting random samples is less time-consuming.

• Stratified Sampling: Stratified sampling refers to the selection of samples from the population that are categorized.

  The samples are divided into strata or layers based on many factors like age, gender, location, and more. Stratified sampling is more time-consuming.

• Convenience Sampling: Samples are selected from a group that is very convenient for the researcher to choose from. No attempts are made for random or stratified sampling. Convenience sampling is best for exploratory surveys. They are easy to collect. It is a quick way to get initial data.

Test your knowledge on Sample Size

Now let's dwell on final thoughts before leaving this page.

Final Thoughts

As I conclude this article, you should have received a holistic view of the survey sample size.

We started by knowing what a survey sample size is. Then we moved on to discuss a good sample size.

We then learned the mathematical formula and terminologies behind sample size calculation, followed by a worked-out example.

We looked into a few use cases of sample size determination. Finally, we closed by discussing types of sampling.

Remember, if you ever find yourself conducting a survey with a very vast target population, then sampling is the way to go. Finding the survey sample size will help you in that case.

Happy Sampling!!!