If you don’t know, use 20000
It’s the total number of people that your sample represents and that reflects the true population (if you don’t know, write 20,000 as the sample size don’t change for target populations larger than that and the difference is not statistically significant).
Margin of error
5% is a common choice
It’s the margin of error you can accept (5% is a common true percentage). When the percentage is smaller, you’re closer to the exact answer at a given confidence level. However for lower margin you require a larger sample size.
Typical choices are 90%, 95%, or 99%
It’s the confidence level you can accept expressed as a percentage (90%, 95% and 99% are common levels). It’s the percentage that expresses how confident you are with the results. For example, 95% confidence interval means that 19 times out of 20 the results lies within the margin of error.
Response distribution (%)
Leave 50% if you’re not sure. In most cases this parameter has to be set at 50%.
The results you expect for each question (standard deviation). If you leave 50%, you go with the most conservative assumption.
The number of respondents needed is:
Minimum recommended size of your sample.