why would the minimum sample size from the new york city data set sample calculation in 2010 be so much bigger than the 20 largest cities sample size calculation in 2000?