Replicability was my first thought as well. These papers make great headlines but how much shaky pyramids of conclusions are built on non-reproducible conclusions?