From my own experiences this is the tip of the iceberg, but the majority isn't fraud per se, more like questionable research practices that cumulatively amount to something similar. So maybe not making up data, but fishing (either variables, people, or models) until you find the right combination. On top of that are all the misuses of things, that aren't really fishing, but rather use of methods that produce significant findings, but for reasons other than what is assumed.

Look up Darrel Huff [1][2]. When there’s enough money on the table, there’s lots of funding (sincere or otherwise) that’s used to try to establish a scientific counter narrative. Stuff like this is more run of the mill corruption than what you see with things like climate change where the underpinnings of the entire industrial economy is at stake (or milder things like Jule).

[1] https://statmodeling.stat.columbia.edu/2012/04/27/how-to-mis... [2] https://en.wikipedia.org/wiki/Darrell_Huff