I'm intending to investigate a factors affecting a small population and I need help with my planning as I'm not well versed in statistics.

I'm looking at a group of students totalling 41000 nationwide, the subset which I intend to investigate makes up 1% of the total. After looking at some ways of analysing it looks like I should use a binary logistic regression, with the 2 Y axis variables being part of subset or not part of subset, and the variables being the different factors affecting them- is that appropriate?

Then I need to work out how many people need to respond to this survey to make it significant, I have looked at sample size calculators but I can't make sense of them, can anyone help me?

Thanks