3 ms·
This is a great example of how macro-statistics are inherently useless to predict individual outcomes (or even individual group outcome). I'd go so far to say
by aresant 14y ago
This is a great example of how macro-statistics are inherently useless to predict individual outcomes (or even individual group outcome).
I'd go so far to say irresponsible to publish without disclaimers since these kinds of studies are so easy to regurgitate and buy into about the "state of the USA".
A glaring example - how can you even consider Japan a "peer" nation to the USA when comparing health?
Ethnicity has been shown, again and again, to have a gigantic impact on life expectancy - so a more compelling study may be USA Japanese with Japanese parents vs. Japanese in Japan.
But then again what about RICH Japanese in USA vs. RICH Japanese in Japan?
Or broken out by Japenese that follow different diets?
Or Japanese that smoke?
Or any of the other "studies" that try to identify individualized trends.
If we're trying to glean anything from this data beyond pageviews for newspapers then segment & compare the data along actually comparable lines - class, ethnicity, and regional lines.
- scarmig 14y agoThat correction might make sense for ethnicity. But class? Defending the USA's health system by saying "well, it has a stark class divide which pre-ordains that a large proportion of its population will get inadequate health care compared to other countries" is only valuable in that it tells us that to improve health outcomes, we've got to break down our class system.
- graeme 14y agoWhen Japanese immigrate to America, their health outcomes quickly converge with those of other Americans. 'Ethnicity' isn't the culprit, usually - more likely culture and national policies.