As is the case every election season, the lamestream media are incessantly trumpetting polls that purport to show that the Democratic Presidential candidate (in this case, Obama) is “winning” the presidential race and holding a huge lead over the Republican candidate. The purpose of this propaganda is clear: to make Republicans think that their electoral effort is a vain one, that they stand no chance of defeating Obama, and that they must just as well give up.
But they are wrong, because the polls are NOT showing Obama “winning the race”. Let’s look at the 8 most recent polls as of today (April 25th, 2012), listed here. The first of these was the Rasmussen poll from April 14th-16th, and the most recent one was conducted by Gallup from April 19th-23rd.
Two of these polls (Rasmussen’s from 4/14-4/16 and Gallup’s from 4/14/-4/18) show Romney leading Obama (by 5 pp and 4 pp, respectively). One other shows Obama and Romney tied at 46% each. But there are 5 polls from this period showing Obama “leading”. So Obama must be leading, right?
Wrong. As Steven Warshawsky pointed out a little less than 4 years ago on AT, there are four problems with polls which should cause you not to take them too seriously. Firstly, polls are often worded in a biased way, which is reflected in their results. Secondly, pollsters almost always adjust polls to give Democrats and Democratic responses greater weight, or select sample groups of voters that are disproportionately Democratic in comparison to actual national party registration. Thirdly, there are many independent voters. And fourthly, the Democratic candidate’s “lead” is usually within the poll’s own margin of error.
Let’s look at these polls. The Economist/YouGov poll from April 14th-April 16th “shows” Obama at 49% and Romney at 42%, but the poll’s MOE is 4 pp. So the correct result could actually be Romney at 46% and Obama at 45%. Very close, but with Romney leading. Similarly, the NBC/WSJ poll from 4/13-4/17 shows Obama at 49% and Romney at 43%, but its MOE is 3.1 pp, so it could actually be Romney at 46.1% and Obama at 45.9% – again, very close, but with Romney leading. And the PPP/DailyKos/SEIU poll that purports to show Obama “leading” Romney 49% to 44% has a MOE of 3.1 pp, so it could actually be Romney at 47.1% and Obama at 45.9%, again, very close, but with Romney leading.
And the other two polls “won” by Obama? The Quinnipiac poll has Obama at 46% and Romney at 42%, and the Gallup poll from 4/19-4/23 purports to show Obama leading Romney by 7 pp, 49% to 42%. The polls’ MOEs are, respectively, 1.9 pp and 3 pp. Ah-ha! Obama is definitely winning the race, because his lead in those polls is beyond the MOE, right? Wrong. Look at the results. According to the Gallup poll, 9%, and according to Quinnipiac, 12% of Americans are undecided. These voters will likely vote for Romney. Certainly there is no reason they will vote for Obama.
Additionally, while this Gallup poll gives Obama a large lead, the previous poll by Gallup – conducted just days earlier, from April 14th to April 18th – showed Romney leading by 5 pp, which suggests that in a matter of 5 days the public shifted in Obama’s favor by a net 12 pp. This is a ludicrous suggestion. Again, to borrow a line from Steven Warshawsky, does anyone know any person who goes through such mental and political gymnastics?
In short, polls are not showing Obama “winning” the race any more than they are showing Romney “winning” the race. The only thing they’re proving is that the race is too close to call and that it won’t be decided until the American people actually step into the voting booths on Election Day. When that day comes, I’m confident the majority of them will pull the lever for Romney.