A psychology lab failed to replicate a memory effect using a stricter preregistered method. The authors wrote that the original effect may be smaller than first reported, context-dependent, or absent under the new conditions.