4 ms·

I agree. They create a really nice combined-source dataset and then get totally lost in some quasi-scientific and quasi-statistical narrative of models, indicat

by wjnc 6y ago

I agree. They create a really nice combined-source dataset and then get totally lost in some quasi-scientific and quasi-statistical narrative of models, indicators, p-values and possible effects that are not there in extended models. On top of it all they don't seem to notice the brittleness of their argument. It only holds for Hispanic women, but not for non-Hispanic whites? How would a pregnancy differ between those groups? I would imagine it doesn't and there are other factors in play. (Hint: little socioeconomic information taken into account, although source of insurance and smoking will pick some of that up.) Without having touched the data I think they show that more flares mean slightly shorter gestation and birth weight. Show that. It's novel. Show it from three angles with different indicators. Prove that it's there. Is it not an artifact? What are other health indicators in those regions doing? Etc. The conclusion reads as if you need to prove your possible wokeness in your tenure track. Sprinkle enough possibly and anything goes, even that which you haven't shown.

klyrs 6y ago
> It only holds for Hispanic women, but not for non-Hispanic whites? How would a pregnancy differ between those groups? If one population lives / works closer to flares than another, they could have greater exposure. If one population works outside more than another, they could have greater exposure. If one population is wealthier than the other, their homes may have more up to date air conditioning, including filtering, leading to lower exposure. If one population is wealthier than the other, women will have greater flexibility to stay home for the duration of their pregnancy > The conclusion reads as if you need to prove your possible wokeness in your tenure track. This reads as if you need to flex your anti-wokeness after ignoring significant environmental differences that white and Hispanic women experience. Some factors I mentioned are confounded -- we've seen lower premature birth rates during the pandemic, and we may have a good opportunity to really understand why so many babies are born premature. This study is a piece of that puzzle.
- jeffbee 6y ago
  I don’t know if this is about woke. It just seems like a thin basis for any result. The birth weight response, in particular. There is as much evidence here that ten nearby oil wells increases birth weight versus 1 nearby well as there is that one well decreased it versus zero.
- wjnc 6y ago
  So shouldn't this be in the paper and preferably in the form of testable hypothesis? I guess we agree that the Hispanic part probably isn't useful unless we imagine genetic differences between the populations? The wokeness alert is in the sloppy use of ethnicity and the use of words like "possibly" without any scientific rigor. If you imagine causal effects going through socioeconomic factors that affect different parts of society differently, do model that explicitly. And in this paper the effect size is that small, that you'll probably end up nowhere. They could have been frank about that instead of a conclusion full of strong words.

by wjnc 6y ago

klyrs 6y ago

> It only holds for Hispanic women, but not for non-Hispanic whites? How would a pregnancy differ between those groups? If one population lives / works closer to flares than another, they could have greater exposure. If one population works outside more than another, they could have greater exposure. If one population is wealthier than the other, their homes may have more up to date air conditioning, including filtering, leading to lower exposure. If one population is wealthier than the other, women will have greater flexibility to stay home for the duration of their pregnancy > The conclusion reads as if you need to prove your possible wokeness in your tenure track. This reads as if you need to flex your anti-wokeness after ignoring significant environmental differences that white and Hispanic women experience. Some factors I mentioned are confounded -- we've seen lower premature birth rates during the pandemic, and we may have a good opportunity to really understand why so many babies are born premature. This study is a piece of that puzzle.

jeffbee 6y ago
I don’t know if this is about woke. It just seems like a thin basis for any result. The birth weight response, in particular. There is as much evidence here that ten nearby oil wells increases birth weight versus 1 nearby well as there is that one well decreased it versus zero.
wjnc 6y ago
So shouldn't this be in the paper and preferably in the form of testable hypothesis? I guess we agree that the Hispanic part probably isn't useful unless we imagine genetic differences between the populations? The wokeness alert is in the sloppy use of ethnicity and the use of words like "possibly" without any scientific rigor. If you imagine causal effects going through socioeconomic factors that affect different parts of society differently, do model that explicitly. And in this paper the effect size is that small, that you'll probably end up nowhere. They could have been frank about that instead of a conclusion full of strong words.