An editorial in the January, 2010 issue of Acta Crystallographica Section E tells the distressing story behind a number of frauds involving papers published in the journal (which is a member of COPE). The fraud is extensive, with apparently at least 70 structures having been shown to be falsified. Although the details of how the fraud was uncovered is scientifically complex, the essence is simple in that it came to light during testing of the checking programs for the journal on previously published papers.
The editors of the journal express, rightly, their shock at such an extensive fraud, but clearly have moved swiftly to correct the scientific record. As they say “It is also noteworthy to point out that the current problems could not have been easily discovered without the availability of the structure-factor files; it will become increasingly important for all journals reporting crystal structures to make sure that they require authors to supply such data in future.” and “It is a strength of crystallography that fraudulent practices can be identified, even retrospectively, by diligent archiving of data and checking”
There are clearly lessons here for other disciplines – without the availability of raw data even very egregious examples of fraud may lie undetected in the published literature.