That number varies greatly by technical field, with engineering disciplines much higher.
A good amount of research in human evolution may fall into this category.
Case-control studies on excavation practices, for example, are very rare.
article at some points reads like a support session for scientists who aren’t feeling the citation love, as it goes through one reason after another that research may be valuable even if no one ever cites it.
Most of the reasons discussed in the article are totally legitimate, and should be part of any conversation about the value of lines of research that don’t provoke lots of additional work on exactly the same model.
But really, I’ve heard that said about humanities papers, not scientific papers, and Van Noorden’s article acknowledges that research in the humanities tends to be more independent, with a higher fraction of research that goes uncited by other workers. The 1990 report noted that 55% of articles published between 19 hadn’t been cited in the 5 years after their publication.