Quote:
Originally Posted by nomadking
Small datasets can't be extrapolated that way. Eg 7 in 1,000 could easily actually mean 4,000 in 1,000,000 or 10,000 in 1,000,000.
|
That would be true if they'd only sampled 1000 people, but actually the 7 in 1000 figure is downscaled from a sample size of hundreds of thousands. So scaling back up is statistically valid.