Optimal sample size

When you perform a survey, the intention is to get a representative image about a number of variables or statements within a certain target group or population. Due to practical reasons (too large, too expensive, too time-consuming,…) it is often difficult to interrogate the total population.  In that case a sample is used. This is a selection of respondents chosen in such a way that they represent the total population as well as possible.

It is very important to use a correct sample size. When your sample is too big, this will lead to unnecessary waste of money and time. On the other hand, when it’s too small, your results will not be statistically significant and you will not come to reliable conclusions.

There are many different sampling methods. One of the most used is the random sample, where all members of the population have equal chances of being selected for the sample.  On the support page of our site is a very useful and easy tool to calculate the minimal sample size needed for a survey conducted on a random sample. The calculation is based on the following parameters :

1. Size of the population
   Here you have to enter the size of the group that has to be represented by the sample. If you conduct an employee survey for instance, your population would be the total staff.  Once the population exceeds 20,000, your sample size will not change very much anymore.
2. Preferred margin of error
   This is the positive or negative deviation you allow on your survey results for the sample, in other words the required precision level.  Suppose in your survey 40% of the respondents pick a certain answer and your margin of error is 2%. This would mean that if you interrogate the total population, you can be sure that between 38% and 42% would pick the same answer.  The smaller the allowed margin of error, the larger your sample will have to be.
3. Desired confidence level
   The confidence level tells you how sure you can be of the margin of error, in other words how often the actual percentage of the population that picks a certain answer, lies within the margin of error. In market research, margins of error are calculated generally for a confidence level of 95%.  This means the survey results will be in line with reality 19 out of 20 times.  If you want a higher confidence level (e.g. 99%) your sample will have to be larger.

Once you have calculated the sample size, you know how many respondents you need to generate.  Then you must still estimate how many individuals out of the population to ask to participate to insure the required number of respondents. For instance, if you send out email invitations and your sample size is 100, and the expected response rate is 20%, then you will have to send out 500 invitations.

After the data-collection phase of your survey you will know the actual number of respondents that have participated.  Unless it happens to be the exact sample size you were looking for, you will then need to calculate the achieved margin of error.

Please note, the confidence level and margin of error calculated by our tool is for a random sample. Furthermore, it assumes the response pattern you receive is normally distributed. For sample sizes above 30, the normal distribution usually will be a good estimation of the actual way the responses are distributed  (see also the central limit theorem). For smaller sample sizes the Student’s t-distribution is more appropriate, but is not supported by our sample size calculator.