They already admitted in their emails that they would rather delete the data before handing it over to anyone else. Well, there you go.
And the obvious reason (well, maybe not so obvious) they're taking this line is because they know that if the original data were made available it would become clear that they had engaged in bald-faced scientific fraud, of the variety that has gotten a number of "scientists" drummed out of their profession, and out of academia, in recent years.
By saying "the dog ate the data," they leave the fraud question open forever (they think) and thereby leave open the possibility that, given the soft-headedness of the academic community when it comes to policing it's own, they'll be able to skate by on tenure and keep their jobs (although some may be saying bye-bye to the nice office with the picture-window view of the Quad and hello to the basement room with the banging steam pipes, next door to the vibration lab).
This may or may not work. I find it very hard to believe that at least some of this data wasn't archived, or at least written down in notebooks. People make back-up copies of things. Some may mysteriously "reappear" or get "undeleted" when the charges and counter-charges start to fly.