This time our friends at Farliegh Dickinson University have put out their latest Public Mind poll and the results are not looking so hot for Governor Romney.
According to FDU, President Obama maintains a 14 point lead over Mitt Romney here in New Jersey. This conflicts with recent polls showing the gap to be much smaller, but is consistent in a sense that it shows Obama with a substantial lead. However, that lead should surprise no one for two reasons:
- This poll was taken over 6 days following the Democrat convention. Some of the gap still represents the post-convention bump that Obama received from President Clinton’s speech.
- Mitt Romney is spending essentially zero dollars in New Jersey and will continue to do so.
This poll essentially proves what I have been saying for nearly a month now. Because of the way that Mitt Romney is running his campaign, focusing nearly all of his funds on seven or eight key states that will decide the election, the national polls are going to give an over representation of President Obama’s meaningful support. To be clear, state by state electoral map polling would not be effected by this problem, but polls that rely on a national sample would.
I wil use the FDU poll as an example. For the sake of argument let’s assume that the FDU poll does not over sample Democrats at 47 percent (which it does). If the sample were more reflective of actual likely voter turnout models, Romney would still show a significant deficit to President Obama.
Why? Because every election has people who change sides and independents/undecideds who make a decision. Romney is not reaching out to a single fence-sitting Democrat this fall. Romney is not making a play for New Jersey’s independent vote. As a result, these people, especially the Democrats, will stick with what they know: President Obama, who is making some sort of push for them in hopes to boost turnout in the Philadelphia media market especially.
If you do not reach out to voters, you cannot expect to win their vote. Thankfully, Romney does not need it. New Jersey (and I would argue Pennsylvania at this point) is a lost cause for Romney. If every Democrat and Republican in New Jersey voted for President Obama, it still would not give him any more than 12 electoral votes. So while the FDU poll may in fact be very accurate for New Jersey, using the same model at the national level creates a problem.
By focusing on 7 states, Romney can put his money where it counts, where the election will be won. However, that means that 43 states where he is not really playing are going to skew the polls against him. Polls that sample the entire nation will be diluted by the fact that respondents who will not be casting deciding votes are over-represented.
Sure, national polling might give us a snap shot of how the American public feels about the two candidates (divided, obviously), but that does not make it a good predictor of the outcome. After all, based on the Romney strategy, it is very possible to win the White House without winning the popular vote.
The Romney strategy requires a change in our time tested polling methods to correct the 2012 fallacy. A change that will not likely be made in time for November 6th.