11/2/26
Bibliometrics - (1) from biblio- = book; metrics = measurements. Measurements of publications, authors, etc. (2) from bibliome = complete corpus of biological texts; metrics = measurements. Measurements of text corpora.
 |  |  |
Agenda and Minutes
- We discussed OpenAlex, Google Scholar, Semantic Scholar, Scopus, and Web of Science. Google Scholar seems the most comprehensive but does not offer an official API, though 3rd party APIs exist. The others seem to have API support. Maybe Semantic Scholar is the 2nd most comprehensive. Scopus and Web of Science seem to have relatively limited comprehensiveness.
- TT will check these out a bit and let us know next time.
- Update on data collection section (MB).
- Update on data analysis section (TT).
- Data analysis section will consider things like author order, last author only, accounting for alphabetical order which could indicate order of contribution or indicate equal contribution (take the midpoint? ... based on the interval midpoint general strategy), interval max strategy, interval min strategy, midpoint vs. geometric mean vs. square proportionality hypothesis, .... There's the whole issue of differences among fields, different meanings for author order among different fields, ....
- IRB application must include a project description and all of our current CITI score sheets attached.
- Other papers that we could check in the future:
- lit review insights
- Egghe's “Hm-index”: divides citations by N for N coauthors. This adjusts for the overall weight of the paper in relation to other papers, but makes an arbitrary unsupported evaluation of each author, that does not account for author order, however problematic that admittedly is.
- "Onjia (2025) proposed a correction formula based on the number of authors and the position of authors" and also the number of pubs the author has overall. They used an arbitrary U shaped position formula. 2025. “Output-Normalized Score (OnS) for Ranking Researchers Based on Number of Publications, Citations, Coauthors, and Author Position.” Publications 13, no. 1: 3. https://doi.org/10.3390/publications13010003.
- ScholarGPS.com: site that gives authors and their adjusted citation indexes. There are various factors going into the index (the strategy of ranking researchers using many factors combined together); number of authors is one, and citations can be divided equally among all authors or all given to each author as user options.
- "Zerem (2017) proposed the Author Contribution Score, a metric for quantifying individual authorship as a fraction of the total credit. In this system, the first author receives the full allocation of points, whereas the corresponding author, if different from the first author, receives half of the total points. The remaining half is distributed equally among all other co-authors. When the first author also serves as the corresponding author, the full point allocation for that role is shared equally among the remaining authors as specified. Notably, their index does not assign any particular importance to the last author in the evaluation."
- Previously, we read another relevant article suggested by TT: https://onlinelibrary.wiley.com/doi/full/10.1111/ejn.70464. The paper gives a method for adjusting author publication counts based on the number of authors of the papers, so more authors gives less credit. They point out that the approach could potentially be used in other metrics (like h-index).
- Previously, we discussed Jain and Chandra (2025), https://arxiv.org/pdf/2503.18236.
- Conclusions:
- Equates last and first authorship - an apples and oranges problem that should be handled with *different* metrics, not one metric that confuses apples and oranges.
- Crediting function is qualitatively "reasonable" at least for some fields but quantitatively it is quite arbitrary. This is a key problem from our perspective. Intervals can do it in a more principled way.
- 2nd-to-last author looks like 2nd author - seems rarely reasonable.
Comments
Post a Comment