What is "agenda propaganda"?

The motivation is mainly just the extreme pressure put on academics to publish relentlessly, combined with a publishing culture that favours "interesting" results. When you set out to test a hypothesis with an experiment, the results could turn out to be more or less remarkable. if the results are unremarkable it's still a valid result and useful data, but it's difficult to report that in a paper and get it published, because the journals favour "interesting" results. So a lot of perfectly valid and useful science can't get published. This creates a big temptation to massage results to create something that looks more striking and will get published. This may start with selectively omitting data or selectively analyzing the data, but it's a bit of a slippery slope towards changing data points that don't support the paper's thesis.

They mention one mitigation scheme in particular that is interesting: in this scheme scientists submit the question their experiment is supposed to answer to the journal before they do the experiment. The journal editors decide whether they want to publish it before they see the experimental results. If they commit to publish, they publish the paper no matter what the experimental results turn out to be. This removes an incentive to massage the data to get published.

There's also the way funding and jobs are allocated: you have to have a real stack of publications to your name and constantly be publishing to keep your job or get your research program funded. This creates an incentive for scientists to pay to be included in the author list of papers they didn't actually work on. And there's the phenomenon of professors putting their name to research by their students where they haven't examined the data in detail, and the fact that there are too many papers for anyone to thoroughly peer review, both of which make it easier to slip manipulated or faked data into publications.

Prestige, money, career. Plus a dash of something else.

They talk about many things, but the one that stands out is incentives. There is a lot of competition in academia, publish or perish, with big prizes, so some people are taking shortcuts.