Updating legacy 'missing' items to data issues
During internal audit work, it became necessary to provide a more accurate list of missing items. This was tricky to do given all the legacy issues for data within Sierra.
During phase 1 physical inventory, macros were used to note if something was found or missing from the shelf. During phase 2 documentation, records with the missing macro would be checked to see if they matched any items recorded as uncatalogued and to check if a record was evidence of a missing physical item or a 'ghost' record i.e. one created accidentally. As part of internal audit work around missing items procedure, these checks needed to be brought forward.
As of November 2024, there were 1111 items with a missing macro but these did not necessarily represent 1111 missing physical items. To support the creation of a more accurate list of missing items, a decision was made to create a new status of 'data issues'. From November 2024, any record given a missing macro could either have its status changed immediately to missing or, if something about the record looked incorrect, it could be given a 'data issues' status with a note saying why this status was given. This change meant that any record with a missing status was highly likely to represent a missing physical item and allow the creation of a more accurate missing item list. Any missing items now recorded in a quarter get passed on to LE&E for further checks.
This still left 1111 legacy item records which needed to be checked as either 'missing' or 'data issues' (although some already had the additional 'withdrawn' status). Each record needed to be checked individually and, in some cases, required items to be checked physically in the stores.
As these items were checked they were given a note:
Status changed to 'data issues' as part of inventory reconciliation work Q3 2025.
Status changed to 'withdrawn' as part of inventory reconciliation work Q3 2025.
Ways to check if a physical item is missing or if the record can be marked as a data issue or withdrawn
Was it overlooked during inventory?
Covers items inventoried pre-2020, including EPB A and EPB American, when the team was less familiar with the data. These records often required a check of a physical item down in the stores.
Scenarios included:
missing a catalogued item bound within a book, especially if the bibliographic records were for two separate items and the item records had identical shelfmarks
missing a catalogued volume which was bound with another volume e.g. volume 2 marked as missing but it was bound together with volume 1
missing item related to an item on the uncatalogued spreadsheet, usually with an incorrect/former shelfmark. These items and records were reconciled and the uncatalogued EPBs spreadsheet was updated
Is there a note on the record?
Is there a note on the bib or item record that suggests a data issue or where an 'missing' item might be.
Examples include:
record made in error - can refer to a cataloguer error or presence of a 'ghost' record. Mark status as 'data issues'
vol 1 only - then a record for a missing volume 2 refers to an item the collection has never held. Mark status as 'data issues'
missing since 1983 - evidence that there was a physical item at some point. Mark status as 'missing'
bound with - evidence that there was a physical item at some point. This is often found on records without a shelfmark and can be used to reconcile records with uncatalogued physical items. Mark status as 'missing' if it cannot be traced
BIB LVL - is the record for part of an item e.g. chapter and article rather than an extra physical item? Data will need to be added to object bib record and the part bib and item record deleted. Mark status as 'data issues'
the note EPB Retrocon in a bibliographic record refers to the the retro-conversion of data in 1999/2000. Be wary of these records as the retro-conversion created a large number of data issues
Is there a digitised image or e-book?
If there is only a single copy recorded and there is a digitised version of it, this is strong evidence of a missing physical copy. When the book was digitised may help track when the item went missing. Be aware, a digitised image or e-book may be of an uncatalogued copy or have been matched incorrectly. Make sure to check the image.
Withdrawn?
Does the record have a note of when it was withdrawn from the collection? If yes, the record needs to have the 'withdrawn' status.
Unfortunately, sometimes items were withdrawn but this was not recorded in the catalogue. This cannot be confirmed without deeper research into the collection files and sales catalogues.
The only scenario where it was simpler to check if an item was withdrawn is if the other copy of the 'missing' item is from the Medical Society of London (MSL) purchase in 1984. For Wellcome to get this collection, they needed to sell any 'duplicate' copies for books where the MSL copy was preferred. The shelfmarks of records where this might have happened was checked against a red ledger in the collection files which contains a list of all the books that were withdrawn and sold during this time.
As part of this missing item work, checks were made using the red ledger and adding notes to records that were already said to be withdrawn e.g. Withdrawn December 1984 as part of MSL deposit sale. The status was changed to 'withdrawn' if there was corroborating evidence on the record e.g. a note found on the shelf saying it was withdrawn. Data issues was used and an extensive update note added to say likely withdrawn if a number was spotted but there was no other corroborating evidence. Sometimes the cataloguers replaced a Wellcome copy with an MSL copy and gave it the same number so additional research will need to be carried out to ensure the record refers to a withdrawn copy rather than a missing replacement copy.
Duplication?
This occurs when multiple bibliographic records have been downloaded/created for the same physical item. This required comparing multiple records to the physical copy.
For example, a cataloguer may have downloaded an incorrect record for a work. When a later cataloguer was searching for the record, they wouldn't be able to find it on the system and so downloaded/created another record. If they didn't notice the incorrect one it remained on the system and a single item would have multiple records. This also occurs when an item's title was Latinised in a record but then a record giving the original language, e.g. Hebrew, was also created.
It may also be a duplicate 'holding' record. This is most likely with boundwiths, especially tract collections, when an item record is incorrectly analytically linked to multiple bibs in the system e.g. EPB/T/564 is a 'holding' record where multiple bibs within the tract have been linked to it (EPB/T/564. 1, .2, .3 etc) but there is a second 'holding' bib where there are no other bib records attached.
Multiple copy? Multi-volume set?
For the first round of checks, you need to assume the record is likely to represent a missing physical item. For example, if the second copy of a book has been found, then there is clear evidence that a first copy, even if not found during inventory, existed in the collection at some point. Whether it was later withdrawn or disposed of will require much deeper research into old physical catalogues, collection files and sales catalogues, which cannot be completed at this time
This is a similar situation to multi-volume sets. Unless there is a specific note on the record, e.g. volume 1 only, it has to be assumed that Wellcome had the full set at one point. So if in a three-volume set, volume 2 has not been found, it is highly likely the record relates to a missing physical item. Unfortunately there is evidence on Sierra of items record being created for volumes, even when there is a note on the record saying Wellcome never owned that specific volume. Once again, this will require much deeper research later before a final decision is made.
Bound item not found in boundwith?
Most likely in the Tracts collection where a shelfmark suggests an item is bound within a certain volume but there is no evidence of it in there and the 'correct' item has already been found. This will need to remain marked as missing as it could be a mis-numbered item and may be in an uncatalogued spreadsheet. Instead an 'Extensive update' note will be added to make it clear it is not part of suggested boundwith.
No collection location
If an EPB does not have a collection location, the record is given the 'data issues' status. These 'locations' e.g. EPB Closed Stores, are gradually being checked as part of other data reconciliation work. It is not worth marking these items as 'missing' as there is no way of finding the item. For these items it is best to wait till the end of the inventory and see what records remain and if any of them can be reconciled. If not, then they can be marked as missing or checked against collection files, card catalogues etc.
Result
Below is a breakdown of how many of the previously 1111 missing items were found and made available, how many are likely to refer to missing physical items, how many records are ‘data issues’ and how many of the items had been withdrawn.
Available
Missing
Data Issues
Withdrawn
49
644
289
129
Everything with a missing macro and a status of data issues also has an extensive update note to show why the record was deemed unlikely to refer to a physical copy. A list of these items can be created to consider any final checks or potential deletion of duplicate records.
In accordance with updated missing items procedure, all records that were suppressed when items were first marked as missing were all de-suppressed. This means records for these items can be seen on the public website, allowing for greater transparency around missing items.
As inventory on the material is not yet complete, it is still possible that items currently marked as 'missing' or 'data issues' will be found. If the status needs to be changed from 'data issues' to 'available', a note should be added 'Status changed to 'available' during inventory checks Qx 202x'. Missing items in the ephemera collection are rare as the majority were never catalogued.
Last updated