Add new comment

Do not overestimate the ability of science to provide absolute judgements. I will use a common, over the counter Covid antigen test as an example of how science actually works. The test kit uses a sensitized test strip in the top half of a foldable card, which must be wetted by the test solution after the loaded swab is inserted into the lower of two holes in the bottom half of the card and folded over.

We think of "negative" and "positive" results, but the test kits are real-world analog, not binary. The lower the viral load in a positive sample (the sample not necessarily the person under test) , the lighter the test line will be. The control line is always full dark if enough solution wets the test strip. A negative test is simply the absence of an line whatsoever, in a positive test both the control and test lines appear, but the test line darkens in response to something binding to viral proteins, whose concentraton in the sample is a variable. A very light test line might be visible to some people but not to others. Thus false negative rates will vary with the user's visual acuity and even room lighting. A magnifying glass and flashlight can reduce this issue but are not mentioned in the instructions.

As it is with these test kits, so it is with science in general. Confidence comes from repeatability, when something has been tested over and over again, confidence increases. Thus, we can have great confidence that drinking causes car crashes, because so many drinking drivers have crashed at higher rates than non-drinking drivers, decade after decade. This is an example of large sample size, and of "in the wild" study.