Saturday, January 05, 2013

Blogs as a way to deal with bad or fraudulent science

Thanks to a colleague, in addition to the controversy discussed here, I have been following closely the saga of science-fraud.org . That website, the content of which has now been taken down (but remains available through google's cache, if you know what you're doing), had been a clearing house where an anonymous primary poster (assisted by anonymous colleagues) reported on biomedical research papers that had what looked like some serious problems with image manipulation. For context, in that research area, a common tool is a Western Blot, where stained or tagged proteins are identified by the position of bands in digital images of electrophoresis gels. Because this data is basically shown in papers just as an image, it is ripe for manipulation by unscrupulous researchers with photoshop. While it might be ok to crank up the contrast, for example, it is definitely not ok to copy and paste sections of image, or erase inconvenient bands, or duplicate images from paper to paper and pretend that they are from different proteins. That site highlighted many many very suspicious images in the literature, and outright accused a number of bio researchers of fraud - the poster contacted the deans of many of these people, as well as their institutional research integrity officers and the national Office of Research Integrity.

That blogger has now been revealed as Paul Brookes, a scientist at the University of Rochester med center. His anonymity was broken when a very angry individual spammed a large number of relevant researchers with an email claiming the Brookes was responsible for that page, which the angry person (pseudonymous, ironically), claimed was a hate site. Brookes has come under serious legal threats; at issue is whether some of the posts were libelous, since they often went well beyond pointing out suspicious figures and directly accused people, in public and to their institutions and grant agencies, of deliberate research misconduct.

There are several lessons to draw from all this, as has been pointed out both by Brookes (in a post that he took down) and others.

  • There is a very serious problem with image manipulation out there in that community. It involves investigators and labs at some major places (as well as minor ones).
  • Anecdotally, it looks like some journal editors in this field are reluctant to look at this issue as closely as it needs.
  • Science-fraud.org did identify serious problems with a number of papers and prompted ORI investigations that uncovered research misconduct.
  • There does seem to be a communal need for a better means of identifying suspicious papers, and blogs could serve this role. I'm reluctant to take page views as a direct measure of that need, though - a lot of people click on articles about Kim Kardashian's pregnancy and the latest celebrity trainwrecks, but that doesn't mean that there is a need for those articles.
  • There really is no such thing as online anonymity. That's why I've never bothered to obscure my identity.
  • Do not post anything online or email anything that you would be ashamed to see on the front page of the New York Times.
  • Publicly accusing people of fraud is a serious business - it's not something that should be done without a lot of consideration of the consequences. The university's lawyers are also unlikely to protect a faculty member who makes public accusations like that without consulting them first.
  • It does make me wonder whether there are similar issues in the physical sciences on this scale.
On a related note, getting caught up on last year's news, I see that a Columbia economist involved in Freakonomics has somehow ended up with $240k in "lost" research money, and had to write a check to the university for $13k. Wow. I'm pretty sure that I would be fired if I couldn't account for $240k of research funds, and claiming that I just wasn't good at bookkeeping would be considered a laughable response.

All in all, a rather depressing roundup. Remember, it's this kind of fraud and poor behavior that just give ammunition to the anti-science parts of society.






No comments: