Obama

Alex · 08-12-2009, 04:32 PM

Statistical outliers will be more common among smaller populations.

Using a baseball analogy, a batting average of .400 is more likely over 20 at bats than over 200. The best hitter in his baseball might rank way behind the guy who only played two weeks of the season.

That's why we need to know not only how the nimbers are different but also whether those differences a statistically significant. I'm guessing they probably are in this case.

But then there are confounding factors thay are hard to eliminate. West Virgina and Santa Clara county could have identical populations and identical healthcare systems and I suspect you'd still see statistically significant differences in many demographic measurements.

Finally_ of course_ there's the fact that correlation does not equal causation.

08-12-2009, 04:32 PM	#10
Alex . Join Date: Feb 2005 Posts: 13,354	Statistical outliers will be more common among smaller populations. Using a baseball analogy, a batting average of .400 is more likely over 20 at bats than over 200. The best hitter in his baseball might rank way behind the guy who only played two weeks of the season. That's why we need to know not only how the nimbers are different but also whether those differences a statistically significant. I'm guessing they probably are in this case. But then there are confounding factors thay are hard to eliminate. West Virgina and Santa Clara county could have identical populations and identical healthcare systems and I suspect you'd still see statistically significant differences in many demographic measurements. Finally_ of course_ there's the fact that correlation does not equal causation.
	Submit to Quotes