2 ms·

> If darker-skinned people have, on average, lower vitamin D levels, then measuring vitamin D levels proxies for skin color. "Proxies for" doesn't have a scien

> If darker-skinned people have, on average, lower vitamin D levels, then measuring vitamin D levels proxies for skin color. "Proxies for" doesn't have a scientific meaning that I'm aware of, so I can't be sure that I understand what you're saying. If what you're saying is that vitamin D levels are correlated with skin color, true, but irrelevant to this study. If what you're saying is that the (poorly substantiated) correlation between Covid19 and vitamin D can be used to claim a correlation between Covid19 and skin color: no, because it's quite possible that vitamin D levels have no effect on Covid19 in people with darker skin, but a huge effect on Covid19 in people with light skin. You also have to keep in mind that your errors multiply across different correlations: the error margin of correlating Covid 19 with skin color via this method is the error margin of correlating Covid19 with vitamin D and the error margi of correlating vitamin D with skin color. If what you're saying is that the (poorly substantiated) correlation between Covid19 and vitamin D can be used to establish a causal relation between Covid 19 and skin color, no, because in addition to the previous problems, you're ignoring that there at least 10 confounding factors besides skin color which also have correlation with vitamin D levels: 1. Latitude 2. How much time you spend outside 3. Which time you spend outside 4. Clothing 5. Sunscreen 6. Liver function 7. Diet 8. Supplementation 9. Cloud cover over time 10. Season If what you're saying is that this study can be used to establish a causal relation between Covid19 and skin color: in addition to all the previous problems, this study can't even really be used to establish a correlation between vitamin D and Covid19, so do we really need to pile on more logical fallacies than are already present?

JumpCrisscross 6y ago
> "Proxies for" doesn't have a scientific meaning that I'm aware of It's a term, roughly, for intermediate variables [1]. Vitamin D levels are a likely proxy for a host of race-variant socioeconomic factors. [1] https://en.wikipedia.org/wiki/Proxy_(statistics) https://en.wikipedia.org/wiki/Proxy_(statistics)
- kerkeslager 6y ago
  By the definition on that page, vitamin D isn't a proxy for skin color. > In statistics, a proxy or proxy variable is a variable that is not in itself directly relevant, but that serves in place of an unobservable or immeasurable variable. In order for a variable to be a good proxy, it must have a close correlation, not necessarily linear, with the variable of interest. This correlation might be either positive or negative. > Proxy variable must relate to unobserved variable, must correlate with disturbance, and must not correlate with regressors once disturbance is controlled for. Skin color isn't an unobservable or immeasurable variable, and it does correlate with regressors once disturbance is controlled for.
b1daly 6y ago
I believe what they are saying is that if their is a correlation between skin color and vitamin D levels (people with darker skin have less vitamin D) and there is another reason people with darker skin have poor outcomes (let’s say bad healthcare), it would make any conclusion that the vitamin D levels have a causal relationship to covid-19 outcomes less likely. To account for this, you would need to control for skin color.

JumpCrisscross 6y ago

> "Proxies for" doesn't have a scientific meaning that I'm aware of It's a term, roughly, for intermediate variables [1]. Vitamin D levels are a likely proxy for a host of race-variant socioeconomic factors. [1] https://en.wikipedia.org/wiki/Proxy_(statistics) https://en.wikipedia.org/wiki/Proxy_(statistics)

kerkeslager 6y ago
By the definition on that page, vitamin D isn't a proxy for skin color. > In statistics, a proxy or proxy variable is a variable that is not in itself directly relevant, but that serves in place of an unobservable or immeasurable variable. In order for a variable to be a good proxy, it must have a close correlation, not necessarily linear, with the variable of interest. This correlation might be either positive or negative. > Proxy variable must relate to unobserved variable, must correlate with disturbance, and must not correlate with regressors once disturbance is controlled for. Skin color isn't an unobservable or immeasurable variable, and it does correlate with regressors once disturbance is controlled for.

b1daly 6y ago

I believe what they are saying is that if their is a correlation between skin color and vitamin D levels (people with darker skin have less vitamin D) and there is another reason people with darker skin have poor outcomes (let’s say bad healthcare), it would make any conclusion that the vitamin D levels have a causal relationship to covid-19 outcomes less likely. To account for this, you would need to control for skin color.