Wednesday, October 26, 2005
Yeah, right: A new poll out yesterday claims that if a presidential election were held right now, President Bush would lose to "the Democratic candidate."
There's just one problem with this entire formulation -- everyone imagines the "perfect Democrat." That individual looks a little different to each person in that 55 percent and there's no way the entire 55 percent could agree on a single, real-live Democrat.
The CNN story I linked to also contains this curious paragraph -- remember, it's a news story.
In the poll, 42 percent of those interviewed approved of the way the president is handling his job and 55 percent disapproved. In the previous poll, released October 17, 39 percent approved of Bush's job performance -- the lowest number of his presidency -- and 58 percent disapproved.
However, all the numbers are within the poll's sampling error of plus or minus 3 percentage points, so it's possible that the public's opinion has not changed at all.
So, the poll numbers go up, but the news story is quick to suggest that it may just be the margin of error. While I often hear television pundits point out this fact, I can't remember a news story ever dismissing a 3-point rise in a poll as statistically insignificant (even though it may be).
Think of it this way, if Bush's record-low approval numbers had dropped three points -- within the margin of error -- do you think the story would include that last paragraph that I quoted, or would the further drop be the lede?