Convince (lobby) a politician that your research will save the trees/whales/global warming/end starvation/any other fear inducing thing and get funding to bribe (lobby) more politicians to further your “research” until anyone would be a fool to question the science.

I feel like this same story happens every year and people are surprised. I often wonder how many “academic” or “scientific” quotes to not conflate valid research and study) need to happen before it becomes as distrusted as politicians.

Academic scientists aren't lobbyists, and politicians aren't giving us money. The funding institutions set research goals. Those funding institutions (staffed by people who actually understand science) don't conduct the research--they choose the winning research plans to fund to achieve their stated goals (1 in 12 applicants, last time I tried). These bureaucrats write budget justifications for Congress--just as the military, Social Security Administration, VA, etc, do. And frankly, compared to those entities, our $40-something billion of the $6 trillion budget is pathetic.

https://static.nationalpriorities.org/images/charts/2021-cha...

https://www.nationalpriorities.org/budget-basics/federal-bud...

The article mentions that there are two drugs that resulted from this research. One of them failed a trial recently. Nobody knows if this fraud means that the drug never could have worked, or if this was just bad luck. So yes: people do measure real-world impact. It's just that it takes a very long time and there are plenty of confounding factors, since even non-fraudulent drug research can fail.

That's true for the rare case that something turns into a professionally run and controlled trial (incidentally there's evidence for many published clinical trials also being fake). But very little research output is ever tested that way. Most doesn't even claim it could have real world impact in principle.

What you're observing is that most research results don't have much impact. The ones that get replicated are definitionally the ones that matter the most, since they lead to controlled trials. In principle this is a pretty reasonable way to allocate limited resources.