Accuracy is paramount in journalism, including science journalism. But so is independence.
To ensure that sources (the people interviewed) don't have undue influence over the story, journalistic tradition demands that they never be shown a story before it's in print. Unsurprisingly, this practice leads to some real bloopers.
Tradition does let a reporter read back quotes to a source to check for accuracy. But the significance of a quote changes completely depending on its context, and other facts matter too, so this has limited value.
In my experience, science editors aren't always as strict about this rule, since they know how easy it is to mangle complex scientific issues. Physical Review Focus, for example, regularly runs an early draft by the author of the original paper. This has uncovered important goofs before publication, which is much better than getting a nasty note later!
The review also raises problems, of course. In their work, scientists often frame issues in complex ways, with lots of caveats and conditions. They also think in terms of jargon that may be quite precise, but unfamiliar to general readers. I find that most scientists appreciate the simple descriptions that a broader audience needs, but every so often a draft comes back from review with extensive, unreadably complex revisions.
Sometimes you just have to say no.
Meeting reports, like the eBriefings I do for the New York Academy of Sciences, also raise different expectations than pure journalism. The primary purpose is to convey the speaker's ideas and to help nonspecialists to get up to speed, rather than to provide complete and balanced view. So NYAS routinely asks speakers to vet their eBriefings before posting. I have found this process quite instructive for using words precisely.
There are some interesting surprises, too. One of the biggest hazards is dual-use words. Although many jargon words are blatantly technical, others have an everyday use as well, which may differ significantly. In physics, for example, words like "work" and "force" have precise mathematical meanings that their ordinary usage doesn't convey.
In biology, in a synopsis of one symposium at NYAS, I had written that a gene was "linked" to a trait. The speaker found this language unacceptable. Of course, much of science is the study of correlations between disparate phenomena, and English tends to express this clumsily, often in the dreaded passive voice. So I tried to mix it up a bit. But in genetics, linkage describes a specific protocol for establishing correlation, and is definitely not the same as "association." Writers must continually walk the line between smooth language and precision.
The most startling suggestion I've gotten, though, was to replace "the" with "an"! I had covered a very interesting symposium on the surprisingly rich interactions between the immune and nervous systems. The problem was that I had written that a specific nervous activity "activates the immune response."
See the problem? The speaker knew many ways that the immune system responds to stimuli. Saying "the" immune response implied that there was only one. The more accurate description, using "an," acknowledges that the response is one of many.
It was an easy change to make (hey, it even shortened the text by a character!) but it forcefully reminded me that no word is too small to be important.