Don’t miss Diomidis D. Spinellis analysis of the impact factor of CS journals in 2010. One of the trends (more anecdotal than anything but still…) is that the impact factor of software engineeering journals is globally decreasing.
Given the importance that government agencies give to impact factors (something with which I don’t agree but that’s a different story) this trend is for sure something we need to follow carefully.
Another interesting fact from the analysis is to see that the impact factor of some journals bounces a lot. In countries (like Spain) where the government gives “extra points” for publishing in journals classified in the first 33% of their category, I can imagine that researchers must have a very hard time when trying to decide where to send their papers (needless to say that in the end, the decision is based more on these extra-scientific factors than on purely scientific reasons).
CS = Computer Science
Hi,
This is a controversial and interesting topic. I would like to give my personal opinion (and knowing yours), but first some facts:
1st: It is required a way for measuring quality of research.
2nd: The best way to do that is counting cites . Obviously, the more cites a work gets the better work is (I mean this work has more impact in the research community).
3rd: One way to do that is grouping cites for journal and then we have the quality of journals, so if a paper is in a high-cited journal then it is assumed that it will be cited a lot. This is the rationale behind impact factor in JCR lists.
So, if paper A has been published in a journal two years ago with a high impact factor and paper B has been published in a conference two years ago, but paper A has one cite and paper B has hundred of cites, then which paper is better: A or B? Even more, what about paper C that is a technical report available at a researcher homepage with hundreds cites? Is it no real impact?
IMHO, impact journal was very nice ten years ago when it was so difficult or impossible to count the cites that an individual paper gets, so grouping them by journal results in a nice heuristic. But now, with all of those tools for counting references, why do we need a list of journals? If you need to know the quality of a research, then you can use Scopus or Scholar…
What’s your opinion???
Best,
Jose
Another thing that should be counted (though I know it’s almost impossible) is whether a citation is positive or negative. Maybe a paper has several citations but all of them say the paper is crap.
And I agree that there are other things that should taken into account (e.g. popular research protoypes).
In the end I don’t think there is a way to automatically evaluate the quality of the work of a researcher. E.g. when evaluating candidates for a position, we could use numbers to help in the pre-selection but in the end only an independent senior researcher could evaluate the research work of a peer.
Yes, of course I see your point and I totally agree that we need well-prepared experts to evaluate these “numbers”, but the questions would be: Could research become more “2.0”, thus avoiding depending on a list (btw, made by a company) for determining the quality of a paper? So, could the community (including evaluators) assess somehow the quality of a work in the same way that Wikipedia community assess its entries? I mean, could we talk about “research 2.0”?