A comparison of the last 3 presidential elections to this one

[Grummgug]

The last 3 elections have an interesting similarity, and it is echoed again in 2012.

First, here are charts of the these election cycles.





Each cycle is the same in that during the summer of the campaign, the democrat had a lead. Gore led Bush for most of summer 2000. Kerry led Bush for most of summer 2004. Obama led McCain for most of summer 2008.

Likewise, Obama has led Romney for most of summer 2012.

In each case, the republican took control in the fall. Bush took leads in the fall of 2000 and 2004 that he never really surrendered. McCain took a lead in September 2008 that vaporized under the weight of the credit crisis and voters finding Sarah Palin unacceptable.

Why do the polls lean so democrat in the summer? Whatever the reason, this is the 4th election in a row where the democrat has held a tepid summer lead. If this election goes the way of the last 3, Romney will take the lead in the fall. Unless the economy surges, or Ryan is another Palin, he may not ever give it up. Now that Romney has selected Ryan, I will be watching this trend even more closely.

[Deleted]

I feel bad for Nader.

[Kalis]

...voters finding Sarah Palin unacceptable.

That has to be the most diplomatic thing that I have ever read.

[Wells]

Your analysis lacks depth.

[Dukami]

Can we put an asterisk on 2000 and say that Bush didn't win Florida? Rather, the Supreme Court declared him the winner without a recount.

[ptwonline]

Pretty oversimplified analysis that I don't think has much predictive power based on what you've presented.

For example, it doesn't take into account the number of people who had made up their minds or those who could still be swayed. I am not sure what it was in the previous elections, but right now the number who have not made up their minds is very, very small.

I think the biggest factor in this election that is not yet accounted for is not actually who people will vote for or how polls will change, but the effect of all the voter suppression policies being put in place in swing states. Obama could very well have a big polling lead and lose because people can't vote.

[Grummgug]

Pretty oversimplified analysis that I don't think has much predictive power based on what you've presented.

For example, it doesn't take into account the number of people who had made up their minds or those who could still be swayed. I am not sure what it was in the previous elections, but right now the number who have not made up their minds is very, very small.

I think the biggest factor in this election that is not yet accounted for is not actually who people will vote for or how polls will change, but the effect of all the voter suppression policies being put in place in swing states. Obama could very well have a big polling lead and lose because people can't vote.

Gallup had the number of undecideds in July 2000 between Bush and Gore as 1/20, or 5%.
http://www.usatoday.com/news/e98/e2244.htm

Yet the race completely swung around from a big Gore lead that summer to a Bush victory.

The Gallup poll shows about one in 20 undecided when it offers four candidates - Gore, Bush, Reform Party contender Pat Buchanan and Green Party candidate Ralph Nader.

[Wells]

You're basing your argument on a common fallacy when discussing statistics.

Gore's polling numbers don't effect the current ones. The odds he had don't effect the current ones.

Its like the gambler saying "I've rolled 4 6's now, the next one has to be a 6".

[Grummgug]

Can we put an asterisk on 2000 and say that Bush didn't win Florida? Rather, the Supreme Court declared him the winner without a recount.

Bush won by the rules of the election, that is the Electoral college matters, not popular vote.

---------- Post added 2012-08-11 at 05:23 PM ----------

You're basing your argument on a common fallacy when discussing statistics.

Gore's polling numbers don't effect the current ones. The odds he had don't effect the current ones.

Its like the gambler saying "I've rolled 4 6's now, the next one has to be a 6".

Actually, you are basing your argument on the fallacy of the straw man. I never said there was definitely a correlation, as you claim. That's a straw man. I asked why have the polls been behaving the same way each of the past 3 elections? I didn't have an answer, but I am simply noticing the pattern. There may be a reason. Or it may be random. But I don't think we should dismiss it simply because we cannot find an obvious reason.

---------- Post added 2012-08-11 at 05:40 PM ----------

One possible reason I can think of is the marriage gap and summer vacation / activities.

The marriage gap is the fact that married couples, especially ones with children, trend republican much more than single adults. Single adults trend much more democrat. When school is out, married couples with children go on vacation for weeks at a time, or are out doing summer activities with their kids that take lots of time, and perhaps they are much less likely to be home to answer that landline telephone when a pollsters wants to find out who they will vote for.

The result could be that republican voters get undersampled in polls over the summer, compared to single adults with no commitments that can lead them away from home for long periods.

So married families all start returning home in late summer / early fall as school starts up, then they are tied to their house to take care of the kids. So they are suddenly ALWAYS home answering that landline.

This could explain this sudden bump we see for republicans in the fall.

The problem is, this is all speculation. I'm not someone who can test this theory. Looking at the Gore/Bush chart is really interesting though. Bush had the lead until school let out. Then he retook the lead when school started up again.