3 ms·

> Can you provide some context here? While I don't necessarily dispute what you are saying, it is a very broad and general statement. You're allowed to dispute

by bsder 5y ago

> Can you provide some context here? While I don't necessarily dispute what you are saying, it is a very broad and general statement. You're allowed to dispute it. It's not an obvious thing. Let's take asthma as an example as it's the most recent thing I read. When you analyze the data about asthma, it turns out that asthma attacks decreased even more than expected during the lockdowns. If you compare people who were partially working from home versus the office against the results from the lockdowns, asthma decreased more than expected for the people suddenly thrown home. Since you have the people originally working partially from home, you get a read on the fact that general air pollution got better so you can tease the environmental effects out. What it looks like is that some significant chunk of asthma is due to general activity of the immune system due to swimming in a daily soup of pathogens. "Long COVID" is also a good example. "Long COVID" put a spotlight on lasting damage due to viral infections--of course, once we started looking we find that a lot of viruses leave lasting damage. So, general viral exposure isn't really as "safe" as we thought. How do you make sure that "grey water" doesn't have viral contamination?

robomartin 5y ago
While I understand your perspective one has to be extremely careful in deriving conclusions from such data or analysis. The only way to be able to approximate a supportable conclusion would be a double-blind study with randomly chosen and normally distributed individuals in both the control group and the group being studied for the effect in question. Isolating for the variable under study is one of the cardinal rules of this kind of statistical analysis. The simpler statement is the often-quoted, correlation is not causation. These things respond to dozens, hundred, thousands of variables and are difficult to analyze without adherence to strict statistical methods.

by bsder 5y ago

robomartin 5y ago

While I understand your perspective one has to be extremely careful in deriving conclusions from such data or analysis. The only way to be able to approximate a supportable conclusion would be a double-blind study with randomly chosen and normally distributed individuals in both the control group and the group being studied for the effect in question. Isolating for the variable under study is one of the cardinal rules of this kind of statistical analysis. The simpler statement is the often-quoted, correlation is not causation. These things respond to dozens, hundred, thousands of variables and are difficult to analyze without adherence to strict statistical methods.