Earlier quoted context omitted.
> But it's a pointless argument to make when your 1% margin of error includes 70 million people. You say that like it's some huge number. 70 million or not, it's still 1%. Numbers get their significance relatively, not in themselves.
> You say that like it's some huge number. 70 million or not, it's still 1%. Numbers get their significance relatively, not in themselves. Relative to what? I can't believe I'm having this conversation. 70 million people is more than the number of people killed in both world wars. 70 million people is more than the population of the UK or Canada. Are you claiming that the world wars caused negligible deaths, or that…
Everyone here knows that if you make the population arbitrarily large, the 1% sample becomes large too. But can you really argue that being able to represent 99% with a binary spectrum isn't a pretty good approximation? What percentage would be good enough for you? Or are you going to say "99.9% isn't good enough because 7 million is a lot of people. that's more than died in X'?