Read the bottom of page 17 and page 18 of the article I linked to.

Given where it was published (a magazine backed by statisticians associations in three countries) it would have been very damaging to the author's (a professor at Imperial College) reputation to have been sloppy about conclusions.