I think you fundamentally fail to understand what I just said. Proper unbiased random sampling allows you to create sub populations that tend to reflect the characteristics of a larger group, biased populations don’t share that relationship.
“Because some animals hibernate, all animals hibernate” is just as flawed as saying “Because only a small percentage of hibernate, no animals hibernate.” Instead the relationship is “Because some animals exist that hibernate, there exist animals that hibernate.”
“Because some animals hibernate, all animals hibernate” is just as flawed as saying “Because only a small percentage of hibernate, no animals hibernate.” Instead the relationship is “Because some animals exist that hibernate, there exist animals that hibernate.”