Quote:
Originally Posted by 'Hamas' Jenkins
They sampled 200 people on the street in Chelsea, Mass. Sixty-four of them have antibodies, but the presence of antibodies doesn't mean you're no longer contagious. Also, this wasn't a random sample; we know that half of them complained of at least one symptom consistent with COVID over the previous week. This may have made them more likely to volunteer a sample. Moreover, the area they chose is overrepresented in demographics that are more likely to have the virus due to their work and living conditions.
The story also doesn't present any useful information about the test or its specificity, which is extremely important.
There is no doubt that we have undercounted the total number of cases, but I would caution drawing broad inferences from a group statistically much more likely to have contracted COVID, especially when the sample size is neither random nor large.
|
This one seemed a little better set up than some of the others I’ve read. I don’t understand why it’s so ****ing hard to set up a rigid good study