3 ms·
The difference here is that height is also correlated with nutrition and by proxy social class. Whatever that study says, eradicating that type of confounding f
by bigfudge 4y ago
The difference here is that height is also correlated with nutrition and by proxy social class. Whatever that study says, eradicating that type of confounding from statistical associations is hard/impossible.
- chrischen 4y agoIt's not as hard as you think, and the correlation with nutrition and social class is disappearing and even non-existent in many places. See Japan: https://www.japantimes.co.jp/news/2018/02/12/national/science-health/average-height-japanese-born-1980-later-declining-study-finds/ https://www.japantimes.co.jp/news/2018/02/12/national/scienc... There's an increase in low birth rate babies (not due to malnutrition), probably due to increased survivability of smaller more frail babies. The fact that they grow up and lead healthy lives means that their society is indeed selecting for the survivability of smaller people.
- bigfudge 4y agoRight, but selecting for in the present tense (ie past 30 years) is going to have a negligible impact on population statistics. Also, ruling out that kind of confounding -is- hard… as mostly practiced any measurement error in confounders reduces our ability to ‘control for’ them… but most epidemiology and all nutrition research just glosses over this.