Both 90% and 8% could be valid sample sizes
Is sample size a number or a percentage? Shouldn’t percentages be irrelevant?
If I start with 1 million subjects, but need to remove 95% for whatever reason, wouldn’t I still have a sample size of 50 thousand?