Discover our method to estimate how many survey responses you need:
How to determine the sample size for an online survey?
One of the first question you have when you launch a survey is "how many survey responses do I need to be confident in the results?". This information is key to be sure you won't make decisions based on biased data.
To help you answer this question, we looked at the data with gathered from hundreds of companies using Screeb to craft the following formula:
n = (Level of Statistical Confidence^2 * Average Screeb Response Rate * (1-Average Screeb Response Rate))/Error Margin^2)
We determined that the error margin would be 5%.
But this formula would work only if you could be sure all your respondents are random. For in-app surveys, the first persons who will see the survey may not be random and so you need a bigger sample size for your survey to be valid.
Then, we mix n with the actual number of people you want to display this survey to to give you an estimation (r):
r = (n / (1+n-1/Audience Size))
How long do I need to display the survey to have enough responses?
Being able to estimate how long your surveys will need to be live before reaching statistical significance is also an important information to size your survey project.
To estimate this, we crafted the following formula:
t = (r/(Audience Size*Average Screeb Response Rate))*24
This way, we can display the number of hours you'll have to wait before being able to rely on your responses.