How can U.S. Census put its name on this nonsense? One needs 10,000 data points to get +/-1% sampling error (see http://en.wikipedia.org/wiki/Margin_of_error ). Are they too lazy to get accurate data on 10,000 data points? 12% margin suggests they collected less than 100 real data points and reported it as a national statistic. Pathetic.
BTW, I have seen similar nonsense in CPI and unemployment rate. For CPI, they conveniently leave out “volatile” food and energy prices. For unemployment rate, they consider anyone unemployed for more than 6 months as “discouraged workers” and leave them out of statistics.
