Final Polling before the election 1980 - note: polls were weighted for the 1976 election image GuardianUK
From Politico: headlines: “Obama's bad October hurts House Dems” suggests that the President is having a George W. Bush effect on his party’s incumbents – apparently there is a loss of confidence in Democrats in general by the public. It is entirely possible, as those casting votes (both early voting and through the election cycle) may choose not only to choose Presidential Candidate, Governor Mitt Romney, over Obama, but will also take a chance on voting against any Democrat on the ticket – similar to what happened in the 2006 and 2008 elections. If history repeats itself, and one takes a look at the 2010 election with a projection of more than less identifying themselves as either Republican or Independent leaning Republican, one can suggest that the 2012 general will look quite a bit like the house trumping of 2010 in a broader sense.
What of the polls? One has to take a strong look at any polling data at this point and basically throw the baby out with the bath water – for example, in 1981, after the landslide victory of Ronald Reagan over former President Jimmy Carter, the head of Gallup explained what had happened and why poll models showed a tight race up to the bitter end. It is suggested that the polling organizations had based their models on the last General election, which had shown an increase in those voting Democrat. That increase percentage was added to the 1980 polls, which gave several points to James Carter that simply did not exist. What was missing from the models, as it was explained by Gallup, the fact that the voter identification had taken a turn to the Republican Party in the two years prior to the 1980 General election. Thus rendering models basically useless. Understanding that models today, for the Presidency, as well as those House and Senate races, are using the 2008 statistics to project winners and losers in the 2012 General. All polls are weighted by a minimum of eight points or more in some cases.
Of course turnout will also have an effect, one cannot dispute simple math. Those identifying themselves as Republican or leaning Republican are significantly more enthusiastic about voting, and there are simply more of them. That said more individuals vote in a General election than a Mid-term, which, using the 2010 model would then skew the poll without compensating for population. The only way, at this point, would be to project using as current as possible voter identification models from each of the fifty states to get a more accurate projection. The fact that Gallup has gone to a “likely voter” model suggests that the polling firm is concerned about using the “weighted” method, as it has, along with the rest of the polling firms, been burnt before (1980 for example).
It is, however, more probable that the 2012 midterm election pattern will be more likely pattern to follow, in part, when one looks at the swing states, or any state to get at least an idea of what might be occurring on a national basis. Although one might find the media “hyping” a certain state, while others have automatically been dubbed “safe Democrat” or “safe Republican” one might want to take a look at what happened in that state during the period of 2009-2010.
Looking at Massachusetts for a moment, one finds that in 2009 special election, the polls and pundits, suggested that then candidate Martha Coakley would best Republican Scott Brown by up to 15 points. The fact that Brown won with 5 points (and the dead and missing voting), suggests that Massachusetts voters changed preference for that election. Going into the 2010 election, all house seats in Massachusetts with one exception saw Republican challengers, the general consensus from the Pundits, was an automatic “safe Democrat” with wide margins of wins projected at 75% of the vote. That was not the case how-ever, as those Democrats, many for the first time, were facing challengers who appeared to have a wind at their backs. In other words, they had to stay home, fight and use their war chests – which kept that extra cash in MA, while it was badly needed for Democrats elsewhere. The final results, although the Democrats held their seats, Massachusetts voters sent a message, as the spread was not a 75% margin, but a 2% to 10% margin.
Therefore it is not without some wonderment, that while Elizabeth Warren the Democrat now running against incumbent Massachusetts Republican Scott Brown, is polling a 5 points – looking at the marginals of some polls taken in the Bay State, the most interesting aspect is the geography – pollsters are using a sample of voters from very low population and very Democrat – Left Ideology, while the sample dismisses (10-25%) the larger populated and more right leaning Worcester County. This also artificially gives Warren a non-existent boost. The Warren Campaign has been running back to back advertisements up until last week, when those ads began to thin out a bit, while Brown’s hammering the airwaves. More over the latest ad is a “hoot” to say the least. In menacing terms, the ad suggests that should Scott Brown win in Massachusetts, he might vote with the Republican Party, and in addition, he might be the one who gives the Republicans a majority in the Senate!
Therefore, what Warren ad is saying, there is nothing essentially wrong with Scott Brown now, except he might be the Republican that takes the Senate away from the Democrats! It smacks of desperation.
The video clip of the ad appears below.
If a Democrat is having problems, regardless of inflated polls, in Massachusetts, where it is suggested that he state is hopeless for Republican’s, what one might suggest is that MA may not be automatically “Safe Democrat” for the General election. It was a state that was certainly a surprise in 1980 to both former President Carter, and President Elect Ronald Reagan.
To further add to the debate: Take Wisconsin’s, where the recent recall of Governor Scott Walker, was a blowout for the incumbent and a resounding defeat of his Democrat Challenger. It is why, at this point, the State of Wisconsin appears “up for grabs” even using the automatic 8 points for Democrats.
In Ohio, the same sample is used, as it is nationwide, therefore, the current tie in Ohio, complete with a new Republican State House and Republican Governor courtesy of the Mid-Terms, may not be an actual battleground – the actual battleground may be elsewhere, in states that are not even under consideration, and holding combined electoral votes to rival Ohio.
Of most interest is the Real Clear Politics electoral college map – take into account this is based on polling data, some overly generous toward the incumbent (specifically those university polls that over-sample low population, heavy Democrat areas, while under sampling, high population, more Republican areas, and then weight an additional 8 points on top. What one finds in the individual polls I that there has been a definitive shift in the voter preferences over that time, and even adding those polls into a combined poll method that Real Clear Politics uses, shows a growing trend towards candidate Mitt Romney. (See Map here at realclearpolticis.com - it is an interactive map, which shows the two candidates with states granted by virtue of having been heavily Democrat or Republican in 2008 – those states are automatically added (MA in the mix), then states considered solid in the polls for either candidate, those leaning towards a candidate and those listed as “toss ups”. (Using polling data, both weighted and over an extended period of time.)
Although a fan of historical data, one must consider in projections that it is best to consider historical trends as well. Understanding that there may be some explaining to do by pollsters on the Wed. following the 2012 elections.