There is a lot less fakery in science than poorly-designed studies, misleading endpoints, underdocumented or incorrectly documented methods, and cargo-culting.

There is a lot less fakery in science than poorly-designed studies, misleading endpoints, underdocumented or incorrectly documented methods, and cargo-culting. The success of the process comes from having a good filtration process to sift through this body of work, and the idea that there will always be some people in the system doing actually good work. That said, I have also witnessed plenty of low-level fraud: changing of dates to match documentation, discarding "outlier" samples without justification or even documentation, etc. Definitely enough to totally invalidate a result in some cases.