From: Stevan Harnad <>
Subject: [SIGMETRICS] Metadata indicating full-text deposits for institutional repositories
Date: 30 October, 2012 8:42:53 AM EDT

Adminstrative info for SIGMETRICS (for example unsubscribe):
Bravo to Dirk and others for their efforts to create reliable full-text deposit metrics
for institutional repositories.

Let me just stress, though, that -- for our recent findings on the correlation between
deposit-mandate-strength and deposits -- the fact that not all deposits are full-texts would
work against, not for, detecting a correlation: 

The deposit-mandates are all full-text mandates, not metadata-mandates. Hence 
whatever the baseline ratio of full-text deposits to total deposits, a significant increase
in total deposits with an increase in deposit-mandate strength is far more likely to
be the result of an increase in full-text deposits than of other kinds of 
content, unrelated to what the deposit-mandate is mandating.

That said, it would be splendid if repositories provided clearer and fuller
metadata to indicate full-text deposits (and, in particular, full-text deposits
of peer-reviewed research articles).

Eprints and Dspace developers (and IR managers): Attention!

Stevan Harnad

On 2012-10-30, at 7:56 AM, Dirk Pieper <> wrote:


the correct term for the "Number of documents" line should be "Number of
OAI PMH metadata records", but outside the information professional
world nobody will understand this I suppose. BASE is harvesting OAI
metadata only.

You are adressing a very important question: what is the ratio of OAI
metadata and Open Access full texts?

Of course there are repositories, which provide 100% open access, but on
the other hand we see that repositories are becoming more and more
platforms for exposing the whole publication output of an institution.
So it would be great, if repository managers would use setSPEC
information about open access to documents more often (there are enough
guidelines from DARE, DINI, DRIVER, OpenAire, ...).

We are trying to indicate real open access to documents within BASE
soon, but I fear that we can indicate this information only for a small
portion of the metadata.


Am 30.10.2012 10:45, schrieb Seb Schmoller:
In the chart does the "Number of documents" line represent "full text
records" or "full text records and metadata only records", and if the
latter is there easily extractable data for each? (Apologies if these
terms are not strictly accurate.)
Seb Schmoller

On 30/10/2012 08:14, Dirk Pieper wrote:

BASE has currently indexed 2.356 repositories, which is more than
OpenDoar but less than ROAR have listed.

This page shows the growth since 2004:

Because we administer the BASE repository list every week when updating
the index, we can assure, that there are not so many skeletons in the
BASE index.