On the other hand, if you're picking up the USA Today, the situation isn't quite so optimistic in this headline: "Bush has ways to go to change Iraq views, poll shows." The USA Today Web site sports this header: "Poll: Bush fails to sway public on Iraq." Meanwhile, CNN's Web site headlines the poll (it's a USA Today/CNN/Gallup poll) this way: "Poll: Iraq speeches, election don't help Bush."
So, which is a news consumer to believe? Well, both of them actually. Or neither, whichever you prefer. The reality is both polls were conducted over the same time period, last Friday through Sunday, both interviewed 1,003 adults across the nation and both have the same margin of error, plus or minus 3%. And headlines for both basically reflect the findings inside.
In the CNN/USA Today poll, Bush's approval ratings rose from 37% to 41%. In The Post/ABC poll, they went from 39% to 47%. Other numbers were comparable. Why? Well, rather than completely bore everyone with statistical analysis and sample breakdowns, it's just as important to keep this in mind: Polls are a snapshot of a given time and specific questions, not always an authoritative reflection of broad sentiment. And something like 1 in every 20 polls are simply wrong, statistically speaking. As President Bush is fond of saying, the polls will go up and down and indeed they do.
The press loves polls because the numbers within are something you can see, touch and chew over. And news organizations love trumpeting the findings of their own polls, CBS News included. Polling isn't cheap and it's hard to justify the cost at times unless there's some "eureka" finding or trend they can tout. So what's a news consumer to do? The best way to figure out just what people are thinking is to take a look at a variety of polling, done over time, and look at the broad mosaic, not just one snapshot.