I was a peer reviewer for a paper that I did not find credible. The data were archived after the paper was published and show many curious or impossible (non-integer counts for example) patterns. I believe the data to be falsified and/or fabricated. I have been publishing my finding on my blog.
One figure, an ordination, is impossible to replicate. Most of the points are correctly located, but some, mostly outliers, are omitted or moved. Different points are omitted or moved in the two versions in the manuscripts I reviewed and the published version of this figure.
This apparent tampering make the relationships in the data appear much stronger than is justified.
I want to publish my findings, including the two manuscript versions of the figure as the inconsistencies between them and the published strengthens my suspicions of malpractice.
Does my duty to expose malpractice supersede my duty to maintain peer review confidentiality?
I have taken my concerns to the Editor, who is not interested.
That's a tricky question. I will try to give you my humble advice.
First, document every step and be sure that your statements are true. If you have some suspicion, you can also present that, but state that you only have such suspicion.
Second, write again to the editor telling him/her that in good faith you believe the study is fabricated, and it will undermine the journal credibility. If you do that after publication, the damage to the scholarly literature will be even larger.
Third, tell the editor your plan. For instance, you can tell him/her that if he does not act swiftly within 30 days, you will inform, without divulging too much details, of the situation, his/her insitution, as well as the authors' institutions and the publisher.
Fourth, if the above does not work (it should), then proceed and be ready to defend thoroughly your argument.