Peter Duniho wrote:


No. To support a blanket generalization, one needs to survey the entire
population.



One needs to do no such thing. Not even remotely close. You can
accurately make a blanket generalization of 300 million people with a
sample of less than 1000.