The practice of running a study, looking at your results, and then gathering more data based on those results. This is invalid in most of traditional statistics. Allowing it is one of the main focuses of safe, anytime-valid inference (SAVI).
Optional continuation is made possible via e-values and e-processes.