Open access and openly accessible: a study of scientific publications shared via the internet

Carr and Harnad (1) bring up a good point that, in my study(2), I did
not draw a sharp distinction between the terms “open access” and “open
access publishing”. I apologize if this led to any confusion as to the
NIH’s stance – Carr and Harnad are correct that the NIH has not suggested
any intention to restrict the choice of journals in which authors can
publish (see http://www.earlham.edu/~peters/fos/nihfaq.htm for a thorough
discussion of NIH open access policy). However, the reason I did not draw
a sharp distinction is that depositing one’s research in a publicly
available archive is functionally the same as open access publishing. In
both cases, readers have free (“open”) access to published research. The
modern definition of “publish” certainly encompasses more than just the
printed form of an article, and thus deposition into a publicly accessible
archive is also a form of publishing. It also needs to be noted that
archiving papers on an individual’s website may be different than
depositing them in a centralized archive (e.g. eprints, arxiv) that
explicitly encourages search engine indexing or submits entries to search
engines. Many of these homepage-based articles may not be as accessible,
both because search engine indexing of web pages is incomplete and because
many home pages eventually “decay” over time as people move, retire or
just neglect to resurrect dead web servers.

Chan and Ng (3) point out that the Adobe PDF format was not adopted
by all journals at equal times, and suggests that the increase over time
in PDFs available at non-journal websites that was reported in my study
may instead reflect the gradual adoption of the PDF format as a standard.
However, this could only be the case if one of three things were true: If
either the online accessibility results were pooled from a broad selection
of journals, all article types were being queried, or journals
created/provided PDF reprints to varying degrees every year. The first is
not the case, as the reported trends of increasing availability were for
individual journals. The second is not the case as only papers classified
in MEDLINE as “Journal Articles” were queried (e.g. some journals
initially provide PDFs for journal articles but not news items or
commentaries because of the added work, but may then later decide to also
provide their commentaries as PDFs). The third does not seem likely, as
each journal adopted the use of PDF format as a means of distributing
reprints at a specific point in time. Although this specific time will
indeed vary by journal as noted by Chan and Ng, it is nonetheless a binary
decision: Either a journal is creating/providing PDF reprints for journal
articles or not – it is not likely to be providing PDF reprints to some
authors and not others. Such a technology adoption curve could only be
captured if journal provision of PDF reprints occurred within the middle
of a publication year, and even then the portion of the curve that
reflects this phenomenon would only span a total of two years. Thus, the
long-term increases in the number of reprints (PDFs) from individual
journals, by recency of publication, that are locatable online is not
likely influenced to a significant degree by a technology adoption curve
as suggested by Chan and Ng.

(1) Carr LA, Harnad S “Open Access Misdefined: Open Access Findings
Missed” BMJ rapid response (15 April 2005)

(2) Wren JD. Open access and openly accessible: a study of scientific
publications shared via the internet. BMJ 2005; 330: 1128-31.

(3) Chan C, Ng DK “Technological problems in this study: PDF” BMJ
rapid response (13 May 2005)

Competing interests:
I am the author of the article being referenced