Or the Germany study, which was also rendered invalid by data, obvious in the same dataset, gathered after publication.