Since the first debate, the most dominant theme has been how Romney’s performance completely changed the race and the polls have been swinging strongly in his favor ever since. As such, I’m sure the next round of polls to come out will also get a lot of attention, mainly, to see if President Obama was able to stop that apparent momentum with his performance.
Via Glenn Reynolds, the latest Rasmussen tracking poll shows Romney continuing to press his advantage. Some caveats apply, like the majority of the data was from interviews prior to the previous debate. So we have a couple more days before any real firm interpretations can be made.
But I’d like to point out this, which points out that Romney was actually moving back to parity with President Obama prior to the debate.
There’s been a lot of talk about the polls and whether they’ve been biased in the Presidents favor. That is, up until they weren’t showing the President with a lead. I’m going to propose a simple reason for Romney’s swing in the polls: bad data.
Now, this doesn’t mean that polling outfits were cooking their results. All it means is that prior to late September, all of the polls were worthless because they were not getting an accurate sampling of the electorate. Then, somewhere in late September, that started to change. This comports with the what seems to be conventional wisdom that most Americans don’t start paying attention to political races until Election Day is close at hand. Most Americans are busy earning a living in a field that doesn’t involve paying attention to politics every single day. Further, my guess is many are jaded to the point that they don’t care about the vast majority of the stuff that gets political pundits excited.
Based on this theory, I don’t think it’s a stretch to conclude that it’s difficult, if not impossible, to get a clear picture of what the electorate is thinking before hand. Thus, pollsters do their thing, but they get hangups or people don’t even bother to answer or participate. (Count me among those- I don’t ever participate.) So whatever data they get is unreliable. It was recently reported that participation in these surveys is down to 9%- meaning 1 in 10 people participate in the polls. It used to be closer to 25%. I think this has to have an effect on polling reliability.
Once that corner is turned in late September; however, the sampling data improves because now more people are engaged and now pollsters are able to get a more accurate sampling (i.e.: better data) of the electorate.
Note that this theory does not predict that we’ll always see a big change in the polls come late September, early October. It merely means that the odds of it happening aren’t as unlikely as some may like to think, especially if, on balance, their preferred candidate has a lot of other problems. It also means that the polls aren’t necessarily as event driven as pundits like to think. Note in my second linked piece, the author is at loss for a reason for the move back to Romney, since nothing noteworthy was going on in the time frame he noted. I think this supports my theory and gives us a period of time where the sampling data began improving.
Also note that this theory does not predict who benefits. It could just have easily been President Obama. Thus, I think it blunts the effect of the first debate. My guess is that debate helped Romney to some degree, but did not have the huge effect most pundits are attributing.
A more interesting scenario would have been if Romney’s performance had been less notable. In this case, my theory means we still would have seen a move towards Romney because that’s where the electorate was already leaning. This might seem odd, but remember, I’m talking solely about the data and its reflection of reality.
One thing I do think it means is that, in this case, President Obama has been operating under a false assumption: that he’s had a solid lead. It would appear that, in fact, he really should have been trying to overcome a deficit.