It's not just honesty. It's an epistemological issue. All science gives most people today is words on paper and a requsst for faith in them. It doesn't give actual knowledge.

Further, because the subjects and objects of a study are analytically defined, how do we know they map to reality?

My thought on that question is that if we had definitions that perfectly matched reality, we would also have perfect results. We would see a 1:1 relationship between treatment and result for example. If we see anything not approaching that, we have either have a poor design or poor definitions. Significance testing is retarded in my epistemology.