Science journals are reeling from the Hwang affair (not to mention the ongoing ramifications of the “hockey stick” controversy. Now comes the news that a study that suggested anti-inflammatory drugs reduce the risk of cancer was based on fabricated data. The research was financed by the National Cancer Institute.

The online Times story linked to above misses out large chunks of the print version of the story. Perhaps most important is the final paragraph:

Science, which published Dr. hwang’s reports, has said it will consider improved procedures, like having all authors state their contributions to a published report. This practice was already followed by The Lancet, however, apparently with no effect in preventing the Sudbo case.

The sort of measures being considered by journals, including investigations into financial interests, are no substitute for the process of real science: genuine, detailed checking of the data by independent reviewers, and replicatory studies. At the moment, peer review is all too often cursory and takes far too much on trust while replication takes second place to original research when it comes to finding funding. As this story proves, being funded by bureaucrats rather than the private sector is no bar to making things up. It’s the data that are important, not the funding.

I’ll leave it to others to make the read-across to other issues.