Replicating a result requires hard work and this hard work does little to help a researcher's career. What helps a researcher's career is if he discover something that is surprising and counterintuitive. But such results are even more likely to wither when subjected to scrutiny.
This is presumably a greater problem in the world-exploring sciences than the math-exploring or engineering-exploring sciences. Verifying a correct proof is much easier than generating one, and if an engineering idea works in practice it will get picked up by others in good Darwinian fashion.