EGAP has endorsed the idea of research registration in its statement of principles. The basic idea is that registration of research designs limits the scope for “fishing” and provides a tool for assessing the scope and effects of publication bias. See here and here for discussions.
But while there is growing momentum towards the adoption of registration as a standard, currently there are few options for social scientists to register their designs. Important options include the AEA trial registry (open only to RCTs) and the RIDIE registry (open to experimental and observational studies assessing the impact of development programs). The EGAP registry is providing an unsupervised stopgap function to store designs until the creation of a general registry for social science research.
If you get to choose your tests after you see the data, you can get whatever results you like. To see the logic try out this fishy test.