There is a failure here to distinguish between two kinds of creationist. One is the preachers and book-writers, the people who tell "their flock" what to think. The other is the audience.

The latter kind isn't necessairly as invested and dedicated as the former; they just pick up creationism by doing what any student does: "learning" from the people who tell them stuff. These are the ones that make debating creationsim non-hopeless; they can be taught real science just as they could be taught the flim-flam in the first place.

When discussing the issue for them (and the audience for a general-public article will include some of them), I think it's best to go after logic first and foremost, rather than the background facts, when given a creationist claim that involves both kind of problem (bad facts and bad logic). And that doesn't need to leave the fact unaddressed, either; for example: "That's not right, but even if it were, it wouldn't mean what they say it means, and here's why...". So every time you hear a creationist sentence, even if the wrong facts are what stands out the most to you, listen for the steps of logic they're asking their audience to take; the question to as yourself while listening to them then is "Even if it's true, what difference would it make?"

For example, if someone says that nobody's ever seen a new species evolve, the natural response of a lot of people is to start listing the cases in which it has been seen, but it would be more effective to point out that evolution doesn't require new species to be formed so quickly right in front of our eyes, and for that matter it would be a serious challenge AGAINST evolution if it DID happen too easily. This not only corrects a misconception about evolution which creationist arguments rely on, but also takes advantage of the fact that human minds tend to find new logical arguments more convincing than factual claims which contradict prior factual claims; it makes the listener more personally involved in figuring things out for himself/herself instead of passively acting as an information receptacle.