The reasons for the "replication crisis" are manifold.
Among them are the incorrect use and reporting of statistical methods, selective publication of desired results due to faulty incentive and publication structures, or the lack of publication of original data, which complicates control functions of the scientific system. To address these problems, several solutions have been proposed that can be implemented at different levels of the scientific process. These include, for example, publishing all original data, study materials, and analysis code, reinforcing the incorporation of good scientific practices into teaching, and changing incentive structures (such as through awards or recognition of these practices in hiring).
Together these are subsumed as “Open Science”.