Literature Citations in the Internet Era
Bibliographic record
Abstract
J. A. Evans's Report “Electronic publication and the narrowing of science and scholarship” (18 July, p. [395][1]) suggests that (i) the average age of citations to scientific papers dropped over the years as more electronic papers became accessible and (ii) the citations are concentrated on a smaller proportion of papers and journals. Such conclusions are not warranted by Evans's data. ![Figure][2] CREDIT: JUPITER IMAGES To measure the evolution of the average (or median) age of the references contained in papers, one has to look at all the references in all published papers and observe the evolution of their age over time. As we have shown using Thomson Reuters's Web of Science data for the period 1900 to 2004 (for a total of 500 million references in 25 million papers), the average (and median) age of all references began to decrease in 1945 but has increased steadily since the mid-1960s. This trend is visible in all sciences, including the social sciences and the humanities ([1][3], [2][4]). The median age of references in fields of science and engineering moved from 4.5 years in 1955 to more than 7 years in 2004, and in medical sciences it increased from 4.5 to 5.5 during the same period ([1][3]). In fact, Evans's conclusions only reflect a transient phenomenon related to recent access to online publications and to the fact that the method used does not take into account time delays between citation year and publication year. Our data also show that in disciplines in which online access has been available the longest (such as nuclear physics and astrophysics), the age of references declines for a number of years in the 1990s but then increases from 2000 to 2007, the last available year of our data set. We have also measured the concentration of citations (and journals) by three different methods, including the one used by Evans. All three measures clearly show that concentration is in fact declining for papers as well as for journals ([3][5]). Although many factors affect citation practices, two things are clear: Researchers are increasingly relying on older science, and citations are increasingly dispersed across a larger proportion of papers and journals. 1. 1.[↵][6]1. V. Lariviere, 2. E. Archambault, 3. Y. Gingras , J. Am. Soc. Information Sci. Technol. 59, 288 (2008). [OpenUrl][7][CrossRef][8] 2. 2.[↵][9]1. D. Tores-Salinas, 2. H.F. Moed 1. V. Lariviere, 2. E. Archambault, 3. Y. Gingras , in Proceedings of ISSI 2007, D. Tores-Salinas, H.F. Moed, Eds. (CSIC, Madrid, 2007), pp. 449-456. 3. 3.[↵][10]1. V. Lariviere, 2. Y. Gingras, 3. E. Archambault , “The decline in the concentration of citations, 1900–2007” ( ). [1]: /lookup/doi/10.1126/science.1150473 [2]: pending:yes [3]: #ref-1 [4]: #ref-2 [5]: #ref-3 [6]: #xref-ref-1-1 View reference 1. in text [7]: {openurl}?query=rft.jtitle%253DJ.%2BAm.%2BSoc.%2BInformation%2BSci.%2BTechnol.%26rft.volume%253D59%26rft.spage%253D288%26rft.atitle%253DJ%2BAM%2BSOC%2BINFORMATION%2BSCI%2BTECHNOL%26rft_id%253Dinfo%253Adoi%252F10.1002%252Fasi.20744%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [8]: /lookup/external-ref?access_num=10.1002/asi.20744&link_type=DOI [9]: #xref-ref-2-1 View reference 2. in text [10]: #xref-ref-3-1 View reference 3. in text
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.052 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.037 | 0.069 |
| Science and technology studies | 0.003 | 0.002 |
| Scholarly communication | 0.012 | 0.015 |
| Open science | 0.001 | 0.004 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.027 | 0.010 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".