Discussion?

[Quentin L. Burrell]

When she signed off from the abstracting service - for a well-earned holiday or  - Gretchen said that the listserv would still be there for discussion. So here goes, with apologies if I seem a little disrespectful at times, but the aim is to provoke response!

1. For me, one of the major disappointments of SIGMETRICS has been the lack of discussion or debate. There have been a few attempts at encouraging discussion, for instance the power law vs exponential law description/modelling of scientometric phenomenon. But here, a haughty, intemperate and somewhat inaccurate response to Rousseau's very pertinent comment brought all discussion to a close. (I apologise, I really should have joined in at that stage.) Also a very few contributors have offered data and some analysis but these seem not to have stimulated any sort of debate on such things as alternative models for the data or methodology for the fitting of models.

Why is this? Does nobody have ideas? Does nobody have questions? Or are we all protecting our ideas ready for publication? 

2. The way it has developed over the past six months or so, SIGMETRICS seems to be primarily a vehicle for sending out "relevant" abstracts. A non-scientific perusal of these suggests to me that a word-search on the name "Garfield" in the list of references, i.e. a citation search, of possible papers is used as the sole criterion. (Please correct me!)  IMHO, reference to Eugene Garfield's work, valuable though it is, is neither a necessary nor a sufficient condition for relevance to the general field of Sigmetrics (even though I still have some difficulties with this term!). If this is the case, then should we perhaps be seeking a rather wider abstracting search? And how should listmembers contribute - we can't just rely on Gretchen.

3. In case the above all seems rather critical, let me finish with a question which should at least warrant factual response but will hopefully engender some discussion. 

It seems that so many "research" papers currently mentioned have titles like "A citation analysis of papers in X, Y or Z" or "The Impact Factor of journals in A, B or C" which are eerily reminiscent of the situation a couple of decades ago when we had many papers presenting "Bradford's law applied to ..." or "A Bradford analysis of ..." and all leading to nothing. The problem in those early days was that the importance of the time dimension had not been realised - it was thought that Bradford's law (whether or not it was true!) - was the same whether we were looking at a "collection" for one year or ten. (And nobody bothered to check!) Is not the same true for much of the current research? 

Where is the theory of time dependent citation analysis? Is there an accepted stochastic theory? Indeed, is there any successful theory or do we just have a collection of empirical studies? 

If this latter is the case, then what is the real value of citation analysis and can it have any scientifically justifiable role in the policy decision making process?

Over to you.

Quentin Burrell


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20000810/51acb5fa/attachment.html>