[samatman]

I would put this only slightly differently: experimental results should be considered unpublishable without the raw data, and irrelevant until replicated.