The end of that quote is "there is something wrong with the way that you are measuring your data". Which changes the meaning of the quote (without it, it's "value anecdotes over data", with it, it's "use inconsistencies with anecdotes to identify problems with data".
Here, I'll formalize it with statistical language: "the likelihood of the events we sample are inconsistent with the model we've used to explain our data."
