This exercise was inspired by a post on the 'BIBFRAME' list asking whether publishers really did re-use ISBNS. Many catalogers have known the answer to this question for years, but I thought it might be worth 20 or 30 minutes to see what I could find out using a few handy tools.
I started with a file of about 800,000 records from a primarily public library cooperative. I used our PLP software to identify all the records containing duplicate ISBNs. A nice thing about PLP is that it will split the dupes into exact matches and 'near-matches' by applying about 20-25 crosschecks once the records with matching ISBNs have been selected.
I loaded the file of near-matches into MARC Report and composed a simple title list. I started scanning through the list of titles for these duplicates, and whenever I found two titles that were radically different, I searched the ISBN on worldcat. After about 1,000 titles, I stopped, having decided that the answer to the querstion was clear.
In those 1000 records, I found 13 groups of duplicates that might qualify for this problem.
I have posted the screenshots of the 13 worldcat results below:
Ideally we need to check the actual items to see what numbers are actually printed on them.