Dear Jim, Thank you very much for your response.
To date, I haven't noticed any particular effort to increase the number of discussion threads. In fact, I think more noticeable inaccuracies could occur from "test" threads and reproduced threads. I suppose these effects might become less significant as the discussions grow but in this sense I agree that the other indicators might be more reliable. In my opinion, the most robust indicators of subsequent publication would indeed be "views" and "ranking". In fact, the articles with the most discussion could be those which are still works-in-progress or which are more controversial. We monitor the views on the number of unique users who have viewed the article. Therefore, it really should not be possible to manipulate this data. Similarly, authors are not permitted to rank their own article and users can only rank an article once. I hope that this helps to clarify. Best regards, James ________________________________________________________________________ James Weeks Chemistry Preprint Server Coordinator ChemWeb Inc. 84 Theobald's Road London WC1X 8RR United Kingdom Tel: +44 (0) 20 7611 4480 Fax: +44 (0) 20 7611 4301 Email: [email protected] http://preprint.chemweb.com http://www.chemweb.com _______________________________________________________________________ -----Original Message----- From: Jim Till [mailto:[email protected]] Sent: 19 August 2001 12:29 To: [email protected] Subject: Re: Elsevier's ChemWeb Preprint Archive On Fri, 17 Aug 2001, James Weeks wrote [in part]: [jw]> each article that is submitted to the CPS has its own discussion [jw]> group where users can comment on the content of the article. It [jw]> is encouraging that there has been sustained use of these [jw]> discussion groups. For example, 45 of the 282 preprints submitted [jw]> now have more than 3 threads in their discussions. I think it will [jw]> also be interesting to monitor how this feature is used. Of the three indicators now available via CPS (views per preprint, score on a 1-5 scale, and number of discussion threads), it's not clear to me which indicator might be the least susceptible to manipulation. All seem quite vulnerable to such manipulation. If any of these three indicators is found to be a reasonably reliable and valid predictor of subsequent publication in a "brand name" journal, then its vulnerability to attempts at manipulation (e.g. via deliberate attempts to initiate more discussion threads) might become a major concern. I'd be very interested in any comments that you may have about this "manipulation issue". [jw]> It is our intention that the CPS will be [OAI] compliant within the [jw]> next two months. Good news! Jim Till University of Toronto
