I'm co-editor for a niche news site (http://pokerfuse.com) and we strictly do not write any headlines in the "is...?" format. It's a good check before publishing anything - if your headline naturally falls in to this format, it should be a signal that either (a) we shouldn't be writing this story, (b) we should stand by our convictions and write the headline in the affirmative, (c) the story needs more research before publishing. We pass up on some easy pageviews, but we see the upside in more loyal/repeat visitors.
why would you insist on starting with an "Is"? A proper yes or no question can start with "Is", "Are", "Does", "Do", "Have", "Has", "Were", "Did", etc etc. How many of your titles end with a question-mark? It's hard to think of how any of those can be false positives.
There was a bit of a brouhaha about this a few years ago with TV pundit Glenn Beck as well. Beck has a certain style of saying he's "just investigating" questions, where he'll posit some extreme allegation, and wonder out loud why the target of the allegation won't go on record denying it and provide evidence. He'll insist he doesn't even necessarily think the allegation is true, but just wants the questions to be aired and discussed openly, and says refusal to do so is suspicious evasion.
In response, someone started a website wondering aloud why Glenn Beck won't deny allegations that he raped and killed a girl in 1990--- the website author, of course, claims he doesn't believe the allegations, but is just wondering why Beck won't address them publicly, and provide evidence of his whereabouts at the time of the alleged incident. Beck, of course, sued for libel in response...
Whenever I encounter a rhetorical question in a headline or link, I assume the author is either going to reach some foregone and facile answer, or no answer at all.
In this sense, questions as headlines are not only linkbait; they're often the signs of worthless articles, usually consisting of regurgitated common sense, or else poorly distilled coverage of superior source material.
There are occasional exceptions, to be sure. But generally, if I can guess what answer an author is going to get to long before he gets there, he's wasting my time.
I often flag or downvote question-titles and "X ways you're..." self-helpy magazine headlines. As you say, it's an almost perfect indicator of bad writing.
A question headline attracts the reader to read the sub-headline. And that is the purpose of headlines (and copy). To get people to read smoothly until they take out the CC and buy.
Problem is that if you overuse it in a blog (for example), then your writing style will end up souding like an infomercial.
Mind you, if you see yourself using linkbait to gain eyeballs, then your blog is just not worth reading (or advertising in).
Recently, I was in a group, and it came up that we don't know why people join the organization and what they want to get out of it. So I asked, why don't we know these things? This question was met with emotional violence. I suspect that the reaction to interesting questions is a litmus test of the intellectual health of an organization.
Arguably, a natural reaction to a question that evokes a strong feeling is to take that as an attack. (Example: If the world was created in 6000 years, how did these seashell fossils get inside mountaintop rock strata?) Often, this natural reaction cuts off an opportunity to learn something interesting.
A feature isn't 100 percent news. It's interest as well, so a question can be fair game. if you're all feature articles though, you're a "soft" outlet.
These question marks are almost as annoying as the "self-interview" question marks made popular by former Sec. of Defense Rumsfeld.
What is a self-interview? Well, it's when you ask yourself a question then immediately answer it. Are they occurring more now than before? Listening to people on TV news shows and NPR seems to show that more are doing it now.
Why would people do this? Because it makes it sound as if they're answering a question that has been asked -- even if the answer they're giving wasn't actually in response to the actual interviewer.
My pet peeve. This was closely watched at Techcrunch and we mocked other blogs (and our own bloggers) who fell back on it frequently (pronouncing titles with question marks with an inflection).
Using a question mark just means you aren't sure of your position or what you are trying to say. It makes it easier to sit on the fence and not really say anything, playing it safe.
I think most readers prefer a firm opinion or statement, regardless of if they agree with it or not.
Then how does one write about interesting open questions? If the implicit assumption that there aren't any interesting open questions, then I have doubts. Such a field should be a mature industry, and not a field for startups or serious research.
I question whether articles without conclusions are necessary at all, and the situation you describe is one of poor writing. Say something, then create a headline based on that. If you aren't saying anything, then a bad headline is all that's possible.
So you're saying it's never necessary to consider open questions at all? What if the question was written as a conjecture, then the name of the conjecture was used as the title, would that be somehow better? That makes no sense.
What's wrong with marketing an article? People respond well to questions, "Top 5...", and other tried and true methods of "baiting" someone into giving your post a chance.
I don't I would have read your post if it was titled like an academic paper.
If you over-use it, it's not effective but for copywriting purposes, you'll get more clicks. Same with email: even if the subject is not a question, put a ? in it, and you'll get more clicks (it's been A/B tested by a few folks)
If the primary motive for question-mark headlines is link-bait, surely this can be more easily solved by getting your headlines from a better source of news?
edit:
NewsItem.objects.published().count()
>>> 648
NewsItem.objects.published().filter(title__istartswith='is', title__endswith='?').count()
>>> 2
(a couple of stories do have this format, but they seem like the rare appropriate cases)