"Democracy is Teetering"

It's not that easy. I had earlier already linked Pierre Duhem, but the point is better made on this page: https://en.wikipedia.org/wiki/Duhem%E2%80%93Quine_thesis

From there:

The Duhem–Quine thesis argues that no scientific hypothesis is by itself capable of making predictions. Instead, deriving predictions from the hypothesis typically requires background assumptions that several other hypotheses are correct — that an experiment works as predicted, or that previous scientific theory is accurate. (...) Although a bundle of hypotheses (i.e. a hypothesis and its background assumptions) as a whole can be tested against the empirical world and be falsified if it fails the test, the Duhem–Quine thesis says it is impossible to isolate a single hypothesis in the bundle.

If you look at the history of science, progress was everything else than smooth. Old theories were supposedly refuted by some and defended by others. This is possible because "falsifying experiments" can be challenged based on challenging background assumptions, quality of measurement and the like. Both quantum and relativity theory had lots of highly qualified opponents in their time, even Newton had them. Furthermore, as I mentioned already earlier, there is in fact not much replication going on in science, and where it is, it often fails. Some stuff is replicated often and therefore of high reliability, because much in research and engineering relies on it, so people have to replicate it in order to have a proper basis for their own constructions. But this is a small minority of very well established material. It is very hard to reach an even remotely similar degree of reliability with anything new researchers come up with.    

Also it isn't enough to come up with experiments that would falsify a theory, they also need to be constructed in such a way that they could actually distinguish between the theory and alternative theories, i.e., in case they are in line with your theory, they need to go against such alternatives, as otherwise the experiment will not be informative either way. Finding such experiments is much harder.

I was going to post the Ioannidis paper, but someone beat me to it.

Probably it was He Who Should Not Be Named. 

I'm not sure why they posted it, but the point I am making is essentially that most research conducted today is improper science, which is fundamentally why there is such a high risk of it turning out to be wrong and, in effect, leading to the now common impression among laypeople that "science is always changing".

This is not Ioannidis' point.