See my previous post, we crossed.
Analyse a city of 1 million people, divide into research groups, see if the results are significant with a very low margin of error, if not, divide the 1 million people in 10 sections of 100.000 people, repeat statistical analysis, you'll probably recognize the...