Sign in with Google Friend Connect.
The Web Cliffology

Election 2012: How true are the polls in the presidential race?

by C. Moult, Publisher,

The 2012 presidential race has had some of the most up-swings and down-swings in presidential polling history.  The polls have been so volatile that political pundits on both sides of the political spectrum have resorted to political spin to make the case that their candidate is in the lead.

Mr. Romney’s campaign has frequently pointed to huge voter enthusiasm among republicans especially after the first debate in Denver and a surge among independent voters as a guide to why their candidate will win.

On the other hand, president Obama’s campaign has frequently pointed to changing demographics and decent numbers among early voters as a driving force for an Obama victory on November 6.

Thus, the central question of this presidential cycle is how true the predictions are from both campaigns based on the polls.

There are essentially two different ways to look at the presidential polling; one being the national polls and the other being the swing state polls.

When we dive into the national polls, we tend to get a more competitive race where in some polls Mr. Romney seems to be doing quite well and in other polls in this late stage of the race, Mr. Obama seems to be doing quite well nationally just before the election.

The Real Clear Politics average of polls as of noon Monday, November 5th currently shows a 0.4% lead among likely voters nationally in favor of Mr. Obama. This is not a big lead by any means, but it represents an improvement for Mr. Obama from his averages during the aftermath of the first presidential debate in Denver, Colorado where he performed poorly.

The swing state polls are quite competitive as well.  According to the polling averages from many polling organizations such as: Real Clear Politics, Talking Points Memo, Five Thirty Eight and Princeton Election Consortium, Mr. Obama seems to be doing well in enough of the critical swing states to achieve 270 electoral college votes.

Many of the political calculations are centered on why there seem to be such a divergence between the national polls which show a much competitive race between the two candidates and the swing state polls which show a much more favorable race for the president.

One way to look at this is that the Obama campaign had spent much of the summer painting Mr. Romney in the key battleground states as a corporate raider who does not care much for ordinary working people, but who only cares about the rich.

This onslaught of negative advertising in the swing states has left Mr. Romney somewhat scarred and bruised in these states while nationally he has improved in the states where little or no advertising were conducted -- essentially the states that do not matter electorally.

For that reason, polls have shown Mr. Romney improving quite a bit in some southern states where he is already the favorite to win and in some traditional blue states where he is not likely to win such as California, Pennsylvania, Michigan and Minnesota.

Mr. Romney’s improvement in some traditional blue states and his improvement in some of the southern states seem to give a fair assessment of why he seems to have the lead or tie with Mr. Obama in some of the national polls.

However, the problem for Mr. Romney is that this does not necessarily offer up an advantage in the race because the winner of the presidential race will be decided on the Electoral College and will come down to a hand full of key swing states. 

Nonetheless, Mr. Romney’s campaign would attempt to mute this argument by pointing to their lead among independent voters nationally and in some of the swing states as a reason for why they have the advantage.

In many polls, both nationally and in some of the critical swing states, Mr. Romney does have an advantage among voters who identify themselves as independents and often by double digit leads.

Because of this decisive lead, we are left to ask: why is the race essentially a tie nationally?

Frankly, there is a reasonable explanation: voters who classified themselves as independents typically lean to one party or the other and in many cases are not true swing voters.

A study done by Josh Marshall at Talking Points Memo has shown that a good sum of adults who are traditionally republicans now classify themselves as independents while many more democrats still consider themselves as democrats and not independents.  This seems to explain why many polls are showing more democrats in their samples than republicans.

To quantify this, let’s look at the fact that both men (Mr. Obama and Mr. Romney) are essentially tied nationally and square that with the fact that on average pollsters tend to give democrats a +7 edge over republicans in their surveys of likely voters.  Also, bear in mind that the +7 edge that democrats typically have over republicans in likely voter surveys, also equates to the exact edge that republicans typically have over democrats among likely independent voters in many of the recent surveys. Thus, if you keep the +7 democratic samples that show up on average in most likely voter surveys and then factor in the +7 republican lead on average in most recent surveys among likely independent voters then you will see that the race is essentially a tie.

The conclusion is that independent voters may not be the deciding factor in this election.  The election will more than likely come down to who can turn out most of their supporters to vote.  The polls this cycle have been extremely volatile and in such a way that each candidate can point to something in the polls to state why their candidate is in a more commanding position to win the election. For Mr. Romney to win, the national polls and battleground state polls must be biased toward him.  On the other hand, for Mr. Obama to win, the national polls and the battleground state polls in this late stage of the race must be accurate as they continue to show him as the slight favorite to win.