Just a note about a very useful project--"SciMeter"-- that enables free, rapid bibliometric assessment of work in physics but that also (I think) has broader implications.
Dr. Hossenfelder mentioned the following in an email, should anyone be interested in providing feedback about the project. "technical queries which concern the functionality of the website, frontend or backend, should be submitted through the contact form on the website. This way they will go to someone who can actually help. For other matters, eg future developments, other metrics, relevance of the whole idea in general, etc, people can contact me directly [at
[log in to unmask] ]."
(For the record, I have no personal invovlement with this project.)
Comments:
1. Assorted efforts like this might help to challenge the undue sway of impact factors and more generally to shift metrics of quality from the journal level, to the individual article level.
2. It would be good to see the physics societies involved in helping to set holistic criteria and evalution for tenure and promotion. Current t and p criteria, whatever they happen to be at any given institution, in my view have a lot to do with perpetuating the glut of scholarly publishing. This in turn creates demand, which in turn fuels the inflationary pricing spiral of journal costs. The whole drive for scientists to create mile-long publication lists of peer-reviewed articles is part of this malaise. Metrics tied to journal cache (as measured by impact factors--IFs) overlook that these factors are a bad way to measure the quality of a single article, since of course single articles vary greatly in how much they contribute to an overall IF.
3. One suggestion I would make, and it's tied closely to the vision of publishing in my preprint about arXiv publishing at ResearchGate, is to develop and enable metrics that would calculate how many times a researcher's work has been cited in a review article--i.e., literature review, not book review!
It would help if arXiv could provide metadata that would identify (even algorithmically) preprints that have a review quality. Won't be perfect, but still I teach about review article searching routinely in my research skills classes. Maybe authors could be required to label their paper as a review article upon submission to arXiv.
Identification of review articles is already done of course in some of the major A and I databases.
We need a bibliometric that tracks numbers of times that a review article whether in the form of a preprint or a peer-reviewed publication happens to cite a particular article. Aggregate citations of this kind could be tabulated for the work of individual authors.
There could be a helpful feedback here: as argued in the preprint, let's shift the focus of journal publishing from peer review of salami slices, to providing review articles. What better use of researcher genius-hours otherwise devoted to picking apart salami sliced research in the peer review process. A metric of this kind would help further reinforce the importance of review articles within physics.
4. This new metric would be very helpful in assess research quality -- from the standpoint that really important work in physics will be synthesized within the narratives about research developments that review articles provide. Notwithstanding the lag time involved before an article gets a mention in a review article.
5. Where would the dollars come from for further development of arXiv along these lines? How about taking part of the library subscription spend for the bloated journals market and devote part of it to arXiv?
P.S.:
--Clarivate of course is well aware of the abuses of the impact factor. This posting is of course not singling them out for these abuses!
--I don't want to imply that the creators of SciMeter are on board with the vision of publishing mentioned above
--views above are not those of my employer
--
Brian Simboli
Science Librarian
Library and Technology Services
E.W. Fairchild Martindale
Lehigh University
8A East Packer Avenue
Bethlehem, PA 18015-3170
(610) 758-5003 Fax (610) 758-6524
E-mail: [log in to unmask]
Profile & Research guides
