The implication is that it's large paragraphs of solid work stolen from a fellow researcher in the field. But it could just as easily mean 22% of her text was scattered with cliches, idioms and excerpts from legislation (or other standard texts) if someone with an agenda was interpreting the results.
https://en.wikipedia.org/wiki/Sandra_Borch_and_Ingvild_Kjerk...
The 'talk' page of that article is completely empty and not a single one of the citations is in English. For all intents and purposes you're taking everything written there on pure faith because the chance that some uninterested observer has double-checked all of those citations is basically zero.
Nor do the tools have a human level understanding of what has or has not been cited properly.