Arrg. Be wary o’ political polls, me hearty. They’re nay always what they seem, me mate. I be talkin’ about th’ recent survey by me two favorite news services in me favorite port o’ call and school, CBS News, th’ New York Times, an’ Quinnipiac University, me bucko.
And one more thing, ya lily livered scurvy dog. If ye use numbers t’ lie, ye be walkin’ the plank.
Can you tell it is International Talk like a Pirate Day, September 19? Let me move my pirate duck out of the way and get my cockatiel dressed like a parrot off my shoulder so I may begin.
Back to the survey. In cooperation with CBS News and the New York times, Quinnipiac University conducted a survey in the “swing states” of Colorado, Virginia and Wisconsin. Below is their summary of the result. The full report can be read at here.
This and other polls are showing the President in the lead, some 45 days from the election. However, some of the numbers are not as exciting as they look. The reason? Standard deviations (Sigma or σ); that plus or minus (±) number that gets thrown around by reporters and statisticians. It is the estimated error from the average.
Now, the calculations for σ are not really important here. What is importation is what it means to you, the voter. In this case, I will refer to the survey’s Colorado numbers.
Doug Schwartz of the Quinnipiac’s Institute of Statistics said that the σ for Colorado is ±2.5 points. That is not a total of two and one-half points, but a five point spread or 2.5 points either side of the average. So, the larger σ, the greater the error. For example;
Questions two and three ask, “Is your opinion of (Barack Obama / Mitt Romney) favorable, unfavorable or haven’t you heard enough about him?”
The favorable response for Obama is 49 percent, for Romney 46 percent which is not an outright win for the President, but a virtual tie.
However, when asked if the respondent favorably supported their candidate (likely voters, questions 4 and 5), the numbers show 73 percent for Obama and 55 percent for Romney. That is significant.
Here is how these numbers will be used. Both sides will use individual questions in their campaign propaganda that support their side or hurt the other. This is what I call the “Rule of the Ridiculous,” using a number so big or so small that it makes your side look the best.
Wal-Mart, the store with the lowest prices, is a fun example. When looking for a new stand mixer last month, we found one in a cooking specialty store for $399.99. (Note it is not an even $400 but $399.99.) Wal-Mart, claiming the lower price, sells the same stand mixer for $399.88. Less money? Yes. Better deal? No.
More important is what the overall numbers tell you? Which numbers are most important?
It is unfortunate that this survey does not provide an overall rating for each of the candidates. It is equally unfortunate that there is no demographic information provided, though one can break that out if you study the entire survey.
However, Questions 1, 1a and 1b are the most telling for me. The survey found two important facts that the news will underreport and the candidates will most likely not discuss. First, though the survey shows overall Obama has the lead, the campaigns sit upon a statistical plateau with 48 percent in favor of Obama v. 47 percent for Romney. Well within the five point spread.
Second, the undecided or not-sure voters make up seven percent of the voting population, not the 30 percent that is usually bantered around. If the election was to be held today and the undecided and not sure are split, there is no way to determine who will win the election. Yes, winning is 50 percent plus one person, but our Constitution does not work that simply. Something about an Electoral College. Damned intellects.
Translated into voterese, most of us have made up our minds. The presidential candidates, as well as those running for U.S. Congress and state offices, will be doubling down their campaigns. Not to focus on changing the minds of the 93 percent, but to capture more than 50 percent of the not-sure-yets.
With less than six weeks until the November 6, a lot of things can happen to change the course of this election. Even the clearest skies can spawn a serious storm that can knock a boat off keel. It will depend on the size of the storm and the ability of the captain to keep his ship right.
So, do nay get excited yet, mates. We do nay know what next high tide’ wi’ brin’ in the morn. Arrrg.