Blute Blog

In their April 2 issue, Nature published an article “on claims in social-science papers put to the test”. It actually was based on three pieces of work on reproducibility, robustness, and replicability. My attention was drawn to the former which found that in hundreds of papers published between 2009 and 2018, the results often could not be independently verified – even using the same methods on the same data. It drew my attention particularly because of the six disciplines studied – economics, political science, psychology, business, sociology, and education – the most precisely reproducible were in that order from most to least. Hmm. Maybe there is a good reason why I have not confined my work over the years exclusively to sociology!