3 ms·
Regardless of whether it's statistically significant in terms of p-value, its pretty hard to discount the possibility the controls might not be effectively appl
by notahacker 7y ago
Regardless of whether it's statistically significant in terms of p-value, its pretty hard to discount the possibility the controls might not be effectively applied (and I find the basic idea that fluoride might have adverse developmental affects on babies highly plausible). When you're comparing fluoridated and non-fluoridated areas of just six cities there are likely to be other regional effects not perfectly captured by the controls - whilst controlling for parental education they found systematic differences in the parents' education levels between fluoridated and non-fluoridated communities. And they attempted that controlling without any data for the most significant factor in child IQ - parental IQ...