I asked some science journalists about this on Twitter. Tight deadlines are a problem, but the bigger issue is that there’s some sort of journalistic principle about not letting “sources” see—-or approve—-the completed article.

I don’t totally understand why, but I think they were a little unclear on what most scientists want, which is more like checking language and details (a lot of words that seem synonymous aren't in technical contexts) than controlling the overall message.