We sometimes get angry emails from authors pointing out that their citation count decreased. And this is quite understandable, as no one likes to see decreasing citation numbers. But it can happen, and most of the time it is for very legitimate reasons, not some conspiracy. Here is a rundown.
- A citing or cited item has been removed from RePEc. This is rare but does happen on occasion. We recommend against removing item in general, as discussed in a previous blog post. Hint: the fact that a working paper got published in a journal is very rarely a reason to remove its link to full text, let alone its entire entry.
- A RePEc archive renumbers its handles. A handle is the identifier of every item in RePEc. It is supposed to be unique and permanent, yet every month some archive has the bright idea to renumber its material. The consequence: the citation analysis needs to be completely redone, and all the other links as well. This leads to a temporary decrease in citation numbers.
- When different versions of some work are linked, citations are consolidated. Thus what counted as separate citations now only counts as one. And this version consolidation can happen for the citing or the cited work.
- The author removed some work from the author profile, for example when a paper got published. Citations to the working paper version are then temporarily lost until the article is indexed in the citation service.
- An error with RePEc. This can happen despite our best efforts, and it always has been temporary.
And lastly, the citation count in the statistics sent to authors is different from what you see on the web at IDEAS or the RePEc Author Services: this is because multiple versions of the same work are aggregated, and because self-citations do not count.
The RePEc Blog 