I love Presidential elections for one because today campaigners do everything by the numbers.
Recent NBC News/Wall Street Journal and Fox/Opinion Dynamics polls show that the two major Democratic candidates would win handily against any of the four major Republican candidates.
But a Gallup poll taken two weeks earlier shows that Rudy Guiliani would beat both Clinton and Obama! All polls were conducted with over 1,000 people sampled from the general population and candidate names were rotated randomly. Could there be that much opinion change in two weeks? Or is there another reason?
One variable seems to be the way the questions were asked. Here is Gallup’s question for Clinton vs. Guiliani:
"Thinking now about the general election for president, which will be held in November 2008. Suppose the election were being held today -- If Clinton were the Democratic Party's candidate and Guiliani were the Republican Party's candidate, who would you vote for: Clinton, the Democrat or Guiliani, the Republican?" If unsure: "As of today, do you lean more toward Clinton, the Democrat or Guiliani, the Republican?"
The NBC News/Wall St. Journal poll question was shorter, and used slightly better grammar:
"If the next election for president were held today, and Guiliani were the Republican candidate and Clinton were the Democratic candidate, for whom would you vote?"
No question can perfectly replicate the same response as a voter in a voting booth, but the discrepancies in answers between the two questions made for as much as an eleven percentage point swing, far beyond the error margin. And did you notice that the person who did better in the poll was the one whose name was mentioned last?
There is one former campaigner who doesn’t seem particularly concerned with poll numbers: President Bush’s approval ratings still hover around 30%.