The data must be reproducible: it means that the data will be reliable and representative and not that this data would lead to one conclusion and another data would lead to another.
Data being varied, unique or surprising does not have anything to do with acceptability of the conclusion.
