Michelle Meyer and Chris Chabris have a long piece in Slate, Why Psychologists’ Food Fight Matters, which covers the argument about replication within psychology. Many readers may have followed the controversy closely, but if you have not then I highly recommend the piece. As Meyer and Chabris suggest this is a general problem in much of science. Arguably it is even more of a problem in biomedical science, since the efficacy (or lack thereof) of drugs and treatments is a more short term life and death matter. But obviously social science is not irrelevant to this world; results can have highly fraught normative consequences.
The general problem is that the ultimate aim of science is to model the world through a process of discovery of robust regularities and then refinement of those general principles. To make an analogy with economic growth theory, the normal workaday science is like classical growth in a Smithian framework. You have parameters which are eventually are maximized (e.g., land), as our knowledge reaches a saturating point with the principles we have at hand. The way to fame, and a critical aspect of science, is what is often termed a “paradigm shift,” which is analogous to the invention of a new technology which changes the fundamental parameters of the game in terms of what seems possible. People find entrepreneurs much more interesting than the CEO of a mature corporation. The latter are fundamentally more like bureaucrats than entrepreneurs, though some can construct a different life narrative. But corporations which execute economies of scale and result in gains in marginal efficiency are a critical part of the ecosystem. Similarly, the corrective and incremental aspects of science are just as important in the grand picture as revolutionary theories and results. But glory goes to the latter.