Reproducibility Salon: what is spooky data

What gives some data a sense of authenticity while other data gets antennae up that something is wrong or "off"? Is spookiness as simple as being "too good to be true" or are there are other features that make a dataset seem wrong? Have you ever investigated a suspicious dataset and found it was real? When did you first feel creeped out by a dataset?

Please RSVP for the address, in Davis Sq., Somerville.