Possibilities I can think of:
-
The data was the result of fudgery/malfunction and has been removed
-
The data is so interesting that the scientists responsible have to verify it before publishing it again
-
The data is so 5D majic that it has to be classified