Popper was just repeating/updating David Hume... and neither has been particularly helpful given how science is actually performed. The most influential writer on the conduct of science was probably Francis Bacon.

Also, only now, with advanced machine learning tools, are we seeing that there are problems involving doctored data in decade-old (and decades-old) manuscripts. So hopefully we can point this out to students and PIs, warning them of consequences to come - loss of reputation and thus funding.

Now that AI is prevalent, it should be possible to determine after a few years if a paper has held up or whether there are significant papers contradicting it - and thereby develop a metric of reliability. Not sure if such a thing would work but that may be becoming possible.

There’s an online iBiology course called ‘Let’s Experiment’, plus books by @steveroyle.bsky.social & @davidjglassmd.bsky.social ‘The Digital Cell’ & ‘Experimental Design for Biologists’ which I found really useful

Also, there are many examples where X is causal of Y but only if Z exists.

Here we say X is necessary for Y but not sufficient to cause Y.

In these instances X and Y may only be correlated when Z exists.