An interesting scientific article from 2013; who can say what the reproducibility is, probably get wildly different results when replicated, especially given the data set is essentially self selected which makes the results bias in many ways.