[mattkrause]

The point is that blank lets you measure that level of "background" contamination, which you then use to correct the measurements made on actual samples.

Suppose you measure around 100 plastic particles per unit in your blank and 1000 in a sample of A. This suggests that A enriches (sheds, etc) microplastic particles. On the other hand, if you found (say) 101 particles/unit in a sample of B, you'd conclude B doesn't do that.

[vel0city]

But in your example you still don't know if it was your testing process that shed 100 plastic particles or if your distilling process shed 100 plastic particles, meaning you don't actually know if B was or was not the source of the plastic particles. Was it your testing process that introduced those 100 particles, was it the distilling process that introduced them, 50/50, or something else?

B would be inconclusive against what you'd hope to be some kind of background, as its not significantly more but one couldn't conclude the source didn't shed that 100 because you don't actually know if in the blank the 100 particles of contamination was definitely your testing process or the source material genuinely having 100 particles of contamination.

I do agree though, in the A case one could pretty easily conclude whatever generated that sample is adding way more particles than an attempt at a baseline/background.

[godelski]

  > B would be inconclusive against what you'd hope to be some kind of background

Correct. And this is why scientists use null hypothesis testing. You disprove things in science, not prove them. I think that's why you're confused. In the first situation you disproved that it comes from the background

[vel0city]

I don't think I'm the one confused, as I'm not saying things like:

> On the other hand, if you found (say) 101 particles/unit in a sample of B, you'd conclude B doesn't do that

I'm fully aware of null hypothesis testing.

[godelski]

I interpreted that sentence as more casual language. Maybe I'm wrong to assume that. But I didn't write it either.

Either way, the difference didn't matter to answer your question. Them getting that part wrong doesn't make the other part wrong nor hard to understand.

  > I'm fully aware of null hypothesis testing.

Great! And now I know you know too