As others have noted, this is a global problem, not just Chinese.
The version that is more difficult to detect is when a cabal of colleagues agree to push each others' papers in this way. So editor A says "you should really quote authors B, C and D." And somewhere else, editor B is saying "you should really quote authors A, C and D."
Machine learning might be a way to tackle this at scale, by teasing out these associations. Of course, this relies on a degree of transparency. Some journals publish all editors' comments and all revisions of a paper. This is a Good Thing, but humans aren't reading all published research, let alone all the meta data.
If someone with relevant ML skills wants to address this, and fancies starting a project, do get in touch :)
A note on the Chinese insinuations that have been mentioned: As always, it's a bit more complex. There may well be reasons that some states might sponsor or 'encourage' gaming of intellectual institutions. If the world is viewed as a zero-sum game, and the currency is power, this unfortunately seems inevitable. Science tends away from this and towards collaboration, but 'politics' often seems to tend toward competition. I've seen university heads explicitly declare to all staff how they intend to game the national rankings, and nobody bats an eyelid, it's business as usual. It's daft and harmful, and frankly I think it requires hard effort from idealistic grassroots activists to address it. Societal improvements are often won through struggle, they're not given away, they don't happen by incremental evolution.
How do you propose to detect if A, B, C and D are a cabal that push their own papers or if they are the people who actually know the subject and want to improve the quality of paper that new people produce?
The version that is more difficult to detect is when a cabal of colleagues agree to push each others' papers in this way. So editor A says "you should really quote authors B, C and D." And somewhere else, editor B is saying "you should really quote authors A, C and D."
Machine learning might be a way to tackle this at scale, by teasing out these associations. Of course, this relies on a degree of transparency. Some journals publish all editors' comments and all revisions of a paper. This is a Good Thing, but humans aren't reading all published research, let alone all the meta data.
If someone with relevant ML skills wants to address this, and fancies starting a project, do get in touch :)
A note on the Chinese insinuations that have been mentioned: As always, it's a bit more complex. There may well be reasons that some states might sponsor or 'encourage' gaming of intellectual institutions. If the world is viewed as a zero-sum game, and the currency is power, this unfortunately seems inevitable. Science tends away from this and towards collaboration, but 'politics' often seems to tend toward competition. I've seen university heads explicitly declare to all staff how they intend to game the national rankings, and nobody bats an eyelid, it's business as usual. It's daft and harmful, and frankly I think it requires hard effort from idealistic grassroots activists to address it. Societal improvements are often won through struggle, they're not given away, they don't happen by incremental evolution.