3 ms·

Medical researches don't, typically. How would that work? "Oh, some pregnant women miscarry after taking this pill? Let's wait to see whether it happens to many

by Arnt 5y ago

Medical researches don't, typically. How would that work? "Oh, some pregnant women miscarry after taking this pill? Let's wait to see whether it happens to many, because correlation≠causation?" Rather, they look for either a mechanism that links the potential cause and observed effect, or for a large enough number to get statistical confidence. And often they issue warnings such as "might want to not give x pill to y kind of people" soon after they start looking, as was done e.g. for the AZ vaccine.

forcry 5y ago
>they look for either a mechanism that links the potential cause and observed effect, or for a large enough number to get statistical confidence From the article. >An analysis of penile tissue from two men receiving penile implants showed the virus was present seven to nine months after their COVID-19 diagnosis. Both men had developed severe erectile dysfunction, Does a sample size of two, look large enough sample to you? That was my point actually.
- Arnt 5y ago
  Two is routinely enough to issue one of those "might"s when there's a plausible mechanism (in this case, the virus is known to cause swelling in other tissue and swelling is known to restrict blood flow in some tissue).
  - forcry 5y ago
    >Two is routinely enough to issue one of those "might"s Sure. "Might" is is a mighty word. A great many things can be accomplished by the appropriate placement of "might"s.