Phase 1: Microform errors
During the retrospective conversion process, many short records from LCS were incorrectly overlaid by full bibliographic records for microform versions of titles we hold in paper. These were isolated and replaced with records for the appropriate paper copies. This phase was completed by July 2000.
- 120 records corrected
Phase 2: Short bibliographic records
This phase consists of short records from the LCS load that had not yet been replaced with full records as part of the retrocon process, often because of typographic errors, orthographic irregularities, or foreign language transcription errors. Work on this phase is ongoing.
- Total short records: 2883
- Total overlaid: 452 (approx. 25% required original, minimal-level records)
- Total remaining: approximately 2100*
This phase has been more complicated than originally anticipated. Records were organized by LC call number to facilitate consultation of the shelflist cards, and the B range includes a disproportionate number of early printed books involving pseudonyms, spurious attributions, bound-with pamphlet collections, and the like, that require extensive bibliographic research and problem-solving. Subsequent call number ranges appear to require less complicated cataloging and language work. Also, we are investigating the use of CatMe to "workaround" network slowness, loss of records in process, and other common problems related to OCLC and network connectivity.
*This number includes 315 Hebrew language records to be handled separately
Phase 3: "Possible errors"
One of the major reasons for undertaking this project was to identify errors in the full bibliographic records loaded for special collections materials during retrospective conversion. Libraries’ staff and researchers have discovered incorrect records caused by "mismatches."
In phase 3, we reviewed a sample of all "retroconned" special collections records, looking for matches on title, publisher, edition, pagination, and date.
- Sample: 1000 records, approximately 12% of total load
- 2.1 % had missing cards or book was not at call number on shelf (at first search)
- 5.5% of sample appears to be a mismatch between the loaded record and the piece (as identified by the shelflist card or the book)
- Total "possible error" rate: approx. 7%
This phase has been by far the most time consuming part of the project because of the number of detailed steps that must be taken in verifying each record. The sample required approximately 80 hours of GAA work. The largest number of mismatches involves discrepancies in paging or volumes (1 v. bound as 2, etc.) A much smaller number have mismatches on editions, sizes, and even titles. As anticipated, records with edition statements or work letters in the call number were much more likely to contain errors. As in phase 2, a subgroup of Hebrew language records (66) were set aside for later resolution.
Phase 3 is currently on hold.
Phase 4: Other Projects
In addition to the 3 original phases of the project, SCCAT has encountered several other types of problems related to database cleanup since July. Completed projects include the creation of item records for serials with full check-in cards, final cleanup of records for materials transferred from RARE’s OSU collection to University Archives, and the like.
We recently learned of approximately 1500 titles in Rare and Charvat still classed with local AH and EL numbers.
Future Plans
While we are committed to correcting errors in the database, the one-year appointments of two GAAs through June 2001 will not allow the completion of the two remaining phases. Therefore, we plan to focus the remaining time of the appointments on continuing Phase 2; some GAA time and staff time will be focused on the AH and EL class numbers in Rare and Charvat (part of Phase 4). Current staffing will allow us to work on isolating those "high-percentage-of-problem" materials from Phase 3 that will require problem-solving only as a lower priority. As always, individual problems brought to our attention are corrected quickly.
We have made great progress towards the goal of an accurate and complete catalog for special collections materials. All the special collections departments have benefited from the variety of database cleanup efforts we have undertaken, as well as from the staffing situation, which allows skilled, trained GAAs to work on this project rather than exclusively reallocating staff from cataloging workflows.
Content last updated:
This page last modified: 3 November 2010