Monday, March 27, 2017

Does making #openaccess publications available as pdf in institutional repositories have a negative effect on stats of publishers version?

Earlier this week I posted about: "Does making #openaccess publications available as pdf in institutional repositories have a negative effect on stats of publishers version?"
I put the topic on my Facebook 


Some researchers are asking librarians to make full-text pdfs of open access in our repository (Pure) invisible on the research portal, because it would affect the read/download stats on the publishers site in a negative way.


In my opinion open access is about making research and publications as easy findable and accessible as possible, and not about boosting the read/download stats for publications on publisher sites..
What is your opinion? What is your library or institutions policy in making the final publishers versions available  as pdf in your repository?

Here are the comments and responses of Peter Suber (in Italic):

Does making #openaccess publications available as pdf in institutional repositories have a negative effect on stats of publishers version? 

Yes. ​Whenever there are OA copies at one site, they will tend to reduce the downloads of the same works at other sites. 

​BTW, while there's pretty good evidence that OA copies in an institutional repository reduce publisher downloads, there's no evidence that they increase publisher cancellations. And on the other side, there's good evidence that they increase citations to the same works.​

Some researchers are asking librarians to make full-text pdfs of open access in our repository (Pure) invisible on the research portal, because it would affect the read/download stats on the publishers site in a negative way. They are talking about the gold oa articles.

​To me, what's remarkable about this is that authors are identifying with the publisher's interests rathe​r than their own interests. Their interest lies in OA, not in protecting the publisher's stats.

Another dimension of this: If the institutional repository is acting with care, then it's only providing OA when it has legal permission to provide OA. The majority of subscription journals provide this permission through their publishing contracts, and the majority of OA journals provide this permission through their open licenses. Hence, the publishers are *consenting* to this OA. There's no reason for faculty to protect publishers when publishers have the means to protect themselves.
In my opinion open access is about making research and publications as easy findable and accessible as possible, and not about boosting the read/download stats for publications on publisher sites..

Exactly. I completely agree.​
Libraries are adding these pdfs partly to build a local archive, just in case anything might happen to the publishers site ( bankruptcy, take-overs etc...)

I agree that repositories also provide the valuable service of preservation. But if anything, providing open access is even more important.


It almost seems as if researchers think that these publications read/download counts, including other altmetric stats, IS going to effect there bibliographic citations.
One fear that is mentioned out loud is, that it would be bad if the repository version (post-print or final versions) would be quoted in a citation. It would cost them "official" citations.

This is another good reason to de-emphasize metrics based on official citation counts!​ 

Altmetrics would not raise this problem, and evaluating scholars by their quality rather than their citation counts would not raise this problem. 

​    
Peter gave a few links to find more evidence quickly:
​I collect some evidence in the updates and supplements to my book (Open Access, MIT Press, 2012, http://bit.ly/oa-book) for pages 152 and 154.

In summary:

Yes, making the pdf of a Open Access article available in an institutional repository, dóes affect the download/read stats on the Publisher's site, bút there is evidence that it increases citations.The permission to make open access available in institutional respositories is legal ánd consented by by Publishers.
The *traffic* is divided between the various locations where open access articles should be offered (personal author site/blog, institutional repository, online Journal....)

No, there is no need to make final publishers versions of open access publications invisible on institutional respositories.
There is a strange fear that other researchers could quote the version available on the institutional respository, but as the actual OA pdf's are downloaded from the Publisher's site, they are always the same for the final publishers versions.
(People might mistake the *green* open access publications in the institutional respository for a citable publication, but when done with care these final author versions are marked and branded very clearly as NOT the final Publishers versions. Many libraries add a special page to the final author version pdf to make this very obvious.)

Yes, also from the preservation aspect, it is a good idea to make the final publishers versions of publications available in the institutional respository.
Post a Comment