I wish I could upvote this comment more than once.
Many reactions in this thread seem to be based on people's pre-existing beliefs, rather than the strength of evidence in the paper itself, which I think is weaker than the article suggests:
I missed your point about the winsorization, and there are many more decisions in the study that could completely change the estimates. This paper is like a case study in the garden of forking paths in research.