I don't where I've heard it, but somebody once said polls are done by media outlets in order to make the news rather than report it. In other words, they use a poll as a way to report on a topic they want to write about, not one in which significant news is not being made. I'm not sure I've ever seen a more obvious case than this:
The AP-mtvU survey found that overall, eight in 10 college students say they feel stress, including four in 10 who say it affects them often. The most often mentioned causes include school, money and relationships. [formatting added]
Headline? "College Students Stressed by War." The paragraph quoted above was the only one out of 27 that didn't address the war. They even managed to include the death toll, number of individual deployments and troops deployed, as well as the number of wounded. My curiosity was piqued, so I went and found the methodology and questions. Out of 21 pages of questions, only four pages covered politics or the war, including opinion questions about PTSD, and veterans' medical care. Note, these questions were being asked of random college students, not veterans or military family members (though there was overlap between the two groups, of course). The other 17 pages of questions addressed the kinds of things you'd expect to be part of a survey about college stress: leisure time, classes, relationships, spring break, drinking, etc.
But of course, the reporter (AP itself?) wasn't really interested in the biggest factors of college stress or how students cope with it. The poll was merely a pivot for a pre-conceived story. Silly me, I always thought newsmakers were people journalists reported on. Turns out "newsmakers" actually work in newsrooms...