e-print e-mbryology

1
e-print e-mbryology
Based on the Los Alamos National Laboratory
Digital Archive, we present our research into the
use of the archive by authors and researchers
from the international physics community.
Professor Stevan Harnad, University of
SouthamptonDr. Les Carr, University of
Southampton Tim Brody, Ian Hickman
Open Citation Project - http//opcit.eprints.org/
2
Growth of the LANL Archive

• Since the start of the archive in 1991 its usage
  has been steadily growing
• Now, after 10 years, it has over 130,000 papers

3
LANL Authors
Number of Identified Authors per Year

• Number of unique names identified in each year
• Before 1995 author meta-data is missing in most
  sub-fields

4
Citation Identification

• Based on automatic extraction and identification
  from the document source (Adobe Acrobat - .pdf)
• We have defined terminology for two types of
  citation
• Red-Link, an author cites a LANL pre-print
  article using a LANL reference (e.g.
  hep-th/0006010)
• Orange-Link, an author cites a post-print,
  published article that is also deposited in the
  archive (e.g. Phys. Rev. D56 6588 (1997))
• Identifies (Red Orange) 600,000 citations of
  3,000,000 total citations from 130,000 papers

5
(No Transcript)
6
Identification Ratio

• Currently 25 of citations are being identified

7
Identification Ratio - hep-th

• Currently 40 of citations from hep-th (High
  Energy Physics - Theory) papers are directly
  citing pre-print articles in LANL

8
Citation Latencies

• The raw data show that the latency of the
  citation peak has been reducing over the period
  of the archive

9
Citation Latencies

• Normalised data are corrupted by an artefact in
  the citation ratios (used to adjust for time)

10
Updates to LANL Pre-prints

• The LANL archive allows authors to update
  articles that they have deposited

11
Update Delay

• There are too few values to provide an accurate
  frequency so a trend must estimated

12
Article Embryology

• Papers with a journal reference J-R cross
  papers without a J-R at an age of 13 months,
  suggesting a time difference of 13 months between
  pre-print and post-print

13
Article State by Sub-field

• Self-professed state of the article (hep is
  updated by SLAC/SPIRES)

14
State of Cited Articles

• Broken down by papers written by authors with
  given impact level
• Author impact determined by Red-Link citations

15
Author Deposit Rates

• 50 of deposits of new papers occur within 4
  months of the authors previous paper

16
Author Impact Analysis

• There is a co-author list for each paper
• Author impact is defined as the number of
  citations an author receives divided by the
  number of papers that author has deposited (the
  mean number of citations for an author)
• By applying this to each author, a list of author
  names with their impact is constructed
• The authors are ranked by their impact
• The set of authors is then divided into three
  impact sets lowest 25, middle 50 and highest
  25.

17
Author Impact Quartiles

• High impact authors update more than medium or
  low
• High and medium impact authors deposit more
  papers than low

18
Authors Papers

• There is no or little occurrence of the single
  high impact paper for the low impact author

19
Citation Spread

• A small number of papers receive a very large
  number of citations