Add ability to prevent CDI content from appearing in Primo, by Alma Electronic Collection
With CDI we have no control over preventing CDI content from appearing in our Primo filtered search, when we have full text holdings which matches to this content.
In many cases this may be a good thing, in not needing to worry over making sure full text content is searchable.
However, it is clear that there are collections in CDI with very poor data, which causes linking to fail, and also has content which we do not want at all such as Journals and Books, as we manage the title level by our Alma records.
For example, CRL Catalog has 76,000 Journal records, which duplicate our own records and many of them are failing Link Resolver requests. LIBRIS has similar issues and content, and so does DataCite.
In just one day of Link Resolver data, for CRL we had 17 of 49 requests fail (34%), and for DataCite we had 27 of 124 requests fail (21%).
These numbers are all real users who are having a poor experience in our Primo environment and building a terrible impression of our Library service.
We should be able to remove this content by Alma Electronic Collection, with this content not appearing in our Primo environment if the CDI Electronic Collection is not in our Alma environment, such as not in our Institution Zone, or a new CDI setting to achieve this.
Please restore our autonomy to remove this content, so that we can offer our users a stable and consistent experience.
Stacey van Groll commented
Hi Dana - This submission is for exactly that enhancement to make it possible for customers to choose metadata sources.
CRL and Datacite are not completely resolved. There are more examples like this, but I'm honestly hesitant to give them because I'd rather focus this submission on the stated premise, rather than having my examples picked apart.
This focus is our autonomy of choice by sites for our users, which we used to have on PCI.
In CDI it is not possible to choose metadata sources separately for discovery. However, the content team is addressing the issues with journal level and other problematic records by removing them from the collections. The issue with DataCite and the CRL catalog are already resolved, others will follow soon. Please note that the activation process for EasyActive customers (as mentioned in the comments in relation to libguides) is something different than what is discussed in this enhancement suggestion
François Renaville commented
Here is another example with CDI HathiTrust collections. None of the CDI HathiTrust collections are activated in our Alma production environment. However, we still get HathiTrust records in search results. I assume they are displayed because of the Match & Merge process, they are considered as records from an alternative collection. This is however very confusing for end-users since (most of the) HathiTrust journal records do have a volume and year indication at the end of the title and that it is most of the time incoherent with access information.
The Yale review. : n.s.,v.1 (Oct 1911-July 1912) (cdi_hathitrust_hathifiles_njp_32101076433117)
View It: Wiley Online Library - Available from 1997 volume: 85 issue: 1.
Journal of the Franklin Institute of the State of Pennsylvania. : v.182 (Jul-Dec 1916) (cdi_hathitrust_hathifiles_njp_32101073435453)
View It: Elsevier ScienceDirect Journals - Available from 01/01/1995 volume: 332 issue: 1.
Additionally, HathiTrust records are flagged as being in Open Access, which is obviously not always the case with the most recent issues being available with a library subscription!
I provided the following scenario in case 00915010: https://www.screencast.com/t/Ur5ns6jAX
1) Our library provides online access to the "Journal of dental research" (ISSN 1544-0591) from vol 78 (1999). SAGE Portfolio is activated in Alma.
2) Beside the Alma record, there are 4 other HathiTrust records: 1 record per year from 1919 to 1922 (but we have NO access to years 1919-1922).
3) When clicking on "Journal of dental research. : v.3 (1921)", View It says that the library provides access from 1999...
--> This does not make sense at all from an end-user's perspective...
Stacey, thank you for submitting this idea. When we do a very general search on something like, "Environmental Studies" and expand the results, 18 of the first 20 records are for LibGuides. For some reason, these records also have a link out to the same Ebook Central Collection portfolio which is not the correct link. There is no link to the actual LibGuides.
We ended up turning this off by default because it is not a good experience for our users and we have no way of improving the accuracy of these search results.
I agree that we can create a better user experience if we have some control over what is included from CDI.
Illinois Wesleyan University
Customization of preferred resources is always an important aspect of shared repositories, as many shared records contain misleading links or pointers. Please do make customization possible for non-selected resources.
Andrea Imre commented
Yes, we are very interested in this as well. There are several libraries that loaded their libguides content into Alma.
If you search the Alma CZ, currently there are 17 collections with "libguides" included in the collection name. The content of these collections pop up in our CDI searches with results that have no relevance to my library's users, sometimes with librarian's names from other institution appearing in the record, without any links - in other words, these are confusing CDI items for any library patron that is NOT from the library that created the libguide. See an example where a librarian's name from Rochester shows up in a CDI record: https://i-share-sic.primo.exlibrisgroup.com/permalink/01CARLI_SIC/12pq4tt/cdi_soarochester01roch_libguide_910934
This information is misleading and completely irrelevant to patrons at any institution other than Rochester. There are over 6800 results when I search Primo expanded article search for "libguides" and filter to "text resources" under resource type.
Giving libraries the choice to deactivate certain collections with irrelevant or in many cases misleading information from CDI searches will improve the user experience in Primo.
Denise Green, CARLI Illinois commented
I agree. I work for a consortium CARLI and libraries have been looking for ways to exclude either certain e-collections from CDI or certain types of material.
Here's some of our notes
For CDI searching, a setting in Alma to exclude certain varieties or collections of Ex Libris’ harvested citations:
1. Selected Institutional Repository Records from other libraries and/or other libraries outside our collaborative network such as:
2. very brief bibs
3. course reserves.
2. Language of publication screening: (if language indicated for textual materials,)
exclude (or send to bottom of relevancy ranking)
materials in languages
the library indicates less relevant.
More control over CDI citations would increase patron satisfaction and save the users time.
For example, LibGuides and other local websites are by their nature rarely helpful to academic library users with no relation to the campus library that sponsors them.
Institutional Repository Records vary wildly in quality and completeness. Some are inherently local such as library LibGuides, on-the-fly bibs, course reserves or ephemera.
Public services librarians know their communities well. They want to improve discovery services by setting what collections are eligible for searching or not. Or if not exclude, then lower in results’ lists by making less relevant.
We realize that most of these questionable results are coming when the "Expand My Results" box is checked but they do appear in other searches too.
Stacey van Groll commented
Your point is a good one, Francois, regarding the sometimes extensive delays in vendor fixes. This is exactly a reason why this design must be changed, to give the option of removal.
We are more than willing to work collaboratively with Ex Libris to fix their index, by opening SalesForce cases. Our evidence of our commitment to this is hundreds of cases open right now.
But I also have a case in for CRL Catalog, and it has been open for 2 months with no response. We should not be forced to present these collections to our users in the interim.
On PCI, we could simply remove these collections from our environment, while we awaited for a fix, with no impact on our users. Now, we must simply just wait for Ex Libris, with literally no option other than removing our own full text holdings, which is of course not an option.
There is also the factor of course that sometimes there may be nothing wrong in some opinions with the collection at all, but we simply don't want it.
This should be our choice, for our environment, for our users.
François Renaville commented
Thanks for submitting this idea, Stacey! It would indeed be very nice if libraries could retain control over what CDI content appears in Primo. CRL Catalog is a real issue for us (#00889287): journal (!) records (in addition to the local Alma records, https://www.screencast.com/t/SggijMgE) and false positive linking (https://www.screencast.com/t/PhNpgdmDZup). Really annoying for a PCI/CDI collection that we have NEVER activated in the last 8 years…
I wish there could be a way for a CDI Inventory Operator to exclude/blacklist some CDI Electronic Collections or some specific CDI records to prevent them to be displayed in Primo. We sometimes report major metadata or linking issues to Ex Libris that rely on vendor response. It can takes months and months before getting a solution (if any). If we could locally take action and hide problematic CDI records (a bit like the Suppress from publishing process), but we would at least be able to locally fix those issues for our patrons.