From a research course in a master's program
This research course told you needed 10% of a population for it to be a representative sample (even if you allow a 3-4% margin of error)? I'd be interested in hearing how they justify that claim.