[nebulous1]

I think you might be mistaken. The study of Nature/Science papers was "Evaluating replicability of social science experiments published in Nature and Science between 2010 and 2015"