Stevan Harnad wrote: > [...] > > (3) Podkletnov actually has an article published in a high-level > journal. It reports a cure for cancer (in reality bogus), involving > drugs that are in reality toxic and do not cure cancer. > [...]
Actually I think the worst possible case is: (5) P. has an article published in a high-level journal. The title and abstract make ambiguous claims for a cure; the paper includes information on toxicity problems with the drug involved, as required by the journal editorial rules (there was a report Monday on NPR of a planned tightening of those rules - apologies for the horrible search URL: http://search1.npr.org/search97cgi/s97_cgi?action=View&VdkVgwKey=%2Fopt%2Fcollections%2Ftorched%2Fatc%2Fdata%5Fatc%2Fseg%5F127007%2Ehtm&DocOffset=14&DocsFound=18&QueryZip=%28Modified+%3D+08%2D06%2D01%29&Collection=ATC&SortSpec=Modified+Desc+Score+Desc&SearchUrl=http%3A%2F%2Fsearch1%2Enpr%2Eorg%2Fsearch97cgi%2Fs97%5Fcgi%3Faction%3DFilterSearch%26QueryZip%3D%2528Modified%2B%253D%2B08%252D06%252D01%2529%26Filter%3Darchive%255Ffilter%252Ehts%26ResultTemplate%3Darchivesearch2%252Ehts%26QueryText%3D%2528Modified%2B%253D%2B08%252D06%252D01%2529%26Collection%3DATC%26SortSpec%3DModified%2BDesc%2BScore%2BDesc%26ResultStart%3D11%26ResultCount%3D10&; from which I'll quote: Rules being drawn up by four of the world's most powerful medical journals will make it more difficult for pharmaceutical companies to manipulate research data on new drugs. Studies published in these journals are extremely influential in persuading doctors of the benefits and safety of new drugs. The medical journals will now insist on guarantees from the authors that they were given complete independence to publish their results, and can personally vouch for the integrity of the data. NPR's Jon Hamilton reports. ) continuing case 5: the drug company implies indirectly that for continued funding P. should "self-archive" a version that does not include the toxicity information. That's just an example of course - all of these are really an issue of who is taking responsibility for the integrity of the information. My belief is that it is important that that responsibility be transferred as far as it can be, from authors to more stable and identifiable entities, so that the end-users of that information have a source for whom they have a certain relationship of trust. "Medscape Select" is one way to do it - though there it will be important to establish the independence and personal integrity of the editor before the outside community is likely to make heavy reliance on it. Note also that "context" can be much more than a table of contents in a journal. An article in Science will sometimes have preceding editorial commentary that may clarify exactly why this article was accepted, and with what caveats it should be read. That sort of thing is almost inevitably lost in an author self-archived database. Even with just a Table of Contents, neighboring articles will often shed other conflicting or supporting light on an interesting new discovery. As I've said before - author self-archives are missing these two elements: a transfer of responsibility, and presentation in a particular context for which the journal seems to be more appropriate. At least in many scholarly fields. > [...] > Self-archived eprints can be designed to carry "health warnings" that > are as shrill as we like a priori (or, more sensibly, a posteriori, > once we get an idea of the size of the bogus paper problem -- if there > is any). With medicine we are talking about lives that can be lost; I don't think a posteriori is good enough if we're seriously hoping that self-archives will be an appropriate means of distributing information to the final practitioners. And if it ISN"T an appropriate means of distributing the information, and the final end-users of the information ignore it in favor of traditional distribution of articles through journals, then where is the motivation for the authors to self-archive (i.e. if they are not reaching any more readers than they otherwise would)? Arthur ([email protected])
