Statistics: Title Counts

See Best Practices for Decision Pages and Tags for groups

Legend: not started IN PROGRESS STALLED decided

Recommendation

Use MMS ID because it is easy and consistent, and because other options will result in undercounting. We will end up overcounting titles where there are multiple bibliographic records for the same thing, especially for electronic resources. However, there is both 1) widespread understanding of the limitations of this approach and 2) widespread acceptance of its utility.

Impact

Background

System-wide, some duplication exists in title-level records. Cataloging rules provide guidance about when separate title-level records are required and when campuses should use the same bibliographic record to represent multiple items. However, in practice, bibliographic records are duplicated with some regularity, especially for electronic materials. For example:

• CDL OA eBook title duplication.
  From CDL e-resources Acquisitions. Summarized email from cdlacq on Sept 21, 2022:

  • CDL turns on full yearly CZ collections.

  • If publishers add some open access titles to either the yearly CZ frontlist collections or give them to us when we receive marc records for our package, then those titles may end up duplicating titles from the same publisher under a full Open Access collection that we have turned on in the NZ.  

  • We determined that this is ok to leave as they are for now, for the sake of staff time & making collections available to users faster in the NZ, compared to the work it would take to clean them up now or prior to processing a new collection.  This is not exactly a new issue. We may consider internally whether there are ways we can try to mitigate this going forward, in the future when we have more capacity for this.

  • Sample: Ex Libris Discovery - 9780806192116 (exlibrisgroup.com)
    
    

    Open

    

• Vendor-supplied, non-provider-neutral, non-OCLC record duplication. Like those from Cassidy cataloging for resources in Westlaw and LexisNexis.

  • 13,700 bibs in UCI Law Westlaw collection

  • 16,650 MMS IDs for Distinct count MMS ID where the Title + Author Combined and Normalized match

  • Example: separate, non-provide-neutral, non-OCLC bib records for every law review in Lexis & Westlaw

• Electronic record duplication when campuses use different record sets. For example, UCI Law is using OCLC records for an eBook package without vendor-supplied records, which is time-consuming but gets us the OCLC updates. But another campus is using non-OCLC CZ records. Both of those sets of records have linked NZ bibs.

• MARCIVE government document records. It’s known that there are duplicative records in the system.

Options Considered

ARL instructions for title counts :

• Report the total number of titles managed and maintained by the library that are cataloged and made ready for use.

• Deduplicate titles by counting multiple copies of the same manifestation as one title.

• Identical content in different formats should not be deduplicated, and each format should be counted as a different title. For example, a serial title available in print, microform and online would be counted as three titles.

• Count different editions and versions of the same work as separate titles since they denote depth in the collection.

• Counting the 245 field when the library provides stewardship for those resources may be sufficient.

Example from UCI Law

At Irvine, Counts from the Titles Subject Area where:

• Lifecycle = “In Repository”

• Resource Type is “Book - Electronic”

• Library Code (Physical) contains LAW OR Library Code (Electronic) contains LAW

Questions to consider

How much concern about duplication exists in the system? In other words, do we already have widespread acceptance for possible overcounting based on shared and longstanding professional expertise about metadata limitations?
Answer: There is some concern but it tends to be context-specific. Systemwide, there does already exist a shared understanding among campus experts that 1) libraries regularly report record counts as an approximation for title counts, and 2) title count data is inherently imprecise.

Action Log