Quote:
Americano
If you ever took a close look at how US GDP and other statistical data is calculated for political feel-good purposes you'd never again trust government numbers. 50k household samplings to determine unemployment for a work force of 135M wouldn't even qualify as a primary test market in the private sector.
|
This statement is either incredibly misleading or incredibly ignorant.
You cannot treat the sampling method for the household survey as though it is for a one-off test market. This survey is done on a MONTHLY basis and were it's sampling size too small to produce relatively accurate and reliable results we would see tremendous shifts on a month-to-month basis regularly. As it is, we have relatively stable data results, indicating sufficient sample size.
For example, if you have a poll with a sample size of X and you take that same poll every week, and each week the results are dramatically different, that indicates an insufficient sample size. If however the results show stability over time, that is an indication of sufficient sample size.
__________________
"It's a good feeling to shoot a bad guy. Something you democrats would never understand. Americans are homesteaders, we want a safe home, keep the money we make, and shoot bad guys!"
----Denny Crane
