Rasmussen Reports does a daily rolling poll on presidential approval and they report the issue in a way I find odd. (All quotes and data from that same cite.)
And then they put up a graphic that shows how the Strongly Disapprove numbers are higher than the Strongly Approve numbers. Today the gap is -13. When I get daily Rasmussen Reports, this is the number they report.
A line later:
This is the metric I expect of presidential approval polling, and this is the one that doesn’t get a graph. (ETA: Er, yes it does. I need to learn how to read. But this graph does come a couple of paragraphs later, so it’s not the more touted graph.)
Why would subtracting the number who Strongly Disapprove from the number who Strongly Approve be valuable, other than as an aside? What does this bit of number crunching actually tell us, practically speaking, that the traditional Approve/Disapprove numbers don’t?
Rasmussen is often accused of massaging their data in a way that supports Republicans. This is the type of thing that gets them that reputation. Since Obama’s opponents will “Strongly disapprove” of, say, any attempt at universal health care, while many of his supporters will only “Somewhat Approve” his particular plan, it exaggerates hostility to the President. That’s why they’re using that metric.
It does somewhat make sense that the “strongly approve/disapprove” should be weighted higher than weak support or disapproval. Although you’d still want to take the latter into account.
The vast majority of political polls are garbage designed to gather particular responses or operate as a ‘push’ to affect the opinion of the pollee. The politicians private pollsters will take the numbers from some of these polls and re-weight them because they understand the manipulation going on.
Old anecdote, can’t verify the source, circa 1900:
Seminary students in a poll were asked if it was allowable to smoke while praying. 100% of the students responded unanimously “no”. When asked if it was allowable to pray while smoking, 100% responded “yes”.
Just like a prosecutor indicting a ham sandwich, a pollster can get any result they want.
I do believe that there is possibly some value to strongly approve vs strongly disapprove. I don’t think it’s as important as general approval vs disapproval, and it’s even less valuable than direct head-to-head polls. 54% of the country may disapprove of Obama, but if 60% disapprove of Romney, that’s going to be the deciding factor in the election.
I love it! Those really are two different questions, although they sound the same. It being allowable to pray while smoking does not mean that it is allowable to smoke in the first place. It merely implies it. And that’s how push pulling works: throw assumptions into the questions, and people who affirmed will defend that assumption, despite it being something they are actually against.