How to find the sample size of your survey?
What is the sample size?
A sample is the set of respondents to be interviewed during a survey. The sample represents a much larger segment of the population that is the subject of the survey study.
Researchers choose their sample often based on specific characteristics such as gender, education level, age or shopping habits, just to name a few.
The sample size is critical to ensure that statistically valid data are collected.
So, it is important to know how to find the right sample size for your survey.
There is a precise formula that can help you calculate the right sample size for your survey.
This formula takes into account 4 factors:
- The population: that is, how many people fall into the demographic group you want to study. For example, for a study about people who drink coffee in Italy, the size of your population will be the total number of people who drink coffee in Italy. Keep in mind that this number is often indicative and not precise.
- The margin of error or confidence interval: that is, how much margin of error you anticipate in your research. This margin is usually expressed as a percentage.
- The confidence level: which is how confident you are that your data will be reliable. This value is also expressed as a percentage and it is usually in line with the margin of error.
- The standard deviation: i.e., an estimate of how much the responses will vary from each other and from the average number. Since at the time of sample creation you have not yet conducted your survey, a safe choice is 50%.
Once you have determined these factors you can proceed with the calculation using a sample size calculator online or paper and a pen.
How to determine your sample size step-by-step
First, you need to identify your z-score (which corresponds to the confidence level) which is the constant value needed for your equation:
Confidence level 90% – Z Score = 1.645
Confidence level 95% – Z Score = 1.96
Confidence level 99% – Z Score = 2.576
These are the most common confidence levels. In case you need different Z Scores, you can find several online tools that can help you!
Now you can use this formula to calculate your sample size:
(Z-score)2 * StdDev*(1-StdDev) / (margin of error)2.
Important note. This equation is only valid over a very large or unknown population.
What if my sample is too small?
In this case you may be interviewing a sample that is not representative of the group you are referring to and may include discrepancies. The data collected is not statistically valid and your research may not be reliable.
What if my sample is too big?
In this case, the project becomes difficult to manage and very expensive, both in terms of costs and resources. The results will be more reliable but the investment will be very heavy.