Lessons from 2010, 2006, & 2002: Why You Shouldn't Read Too Much Into Early Generic Ballot Polling

Democrats are polling slightly ahead of Republicans in the generic ballot. But the GOP is still better positioned than at this point in 2009.

It's been a rough few months for President Obama, both in terms of polling and external political events.

Barely 6 months into his 2nd term, he's seen both his job approval and favorability ratings slide, while unexpected and distracting scandals involving the IRS and the NSA have angered both opponents and allies.

Yet for all the President's troubles, his party has held up well, at least from the perspective of 2014 Midterm Election polling to date.

Excluding Rasmussen Reports, the Republicans have led the Democrats in only one generic House ballot survey taken since the November 2012 election (out of 26 total polls), and have averaged a 43-40% deficit against the party that currently holds the White House and the Senate. See the table below:

*because Rasmussen tracks the generic congressional ballot every week, their numbers in the above chart are monthly averages. Data compiled from Huffington Post Pollster, Real Clear Politics, & TPM Polltracker.

The Democrats' 3 point house ballot advantage seems all the more impressive when considering the fact that the President's party very rarely performs well in second-term, midterm elections.

But try not to draw too many conclusions from the current state of polling regarding the 2014 election, especially considering the state of generic House ballot polling at this point in 2009.

From immediately following the 2008 Presidential election, until August of 2009, the Democrats averaged a House ballot lead over the Republicans of 44-38%; this, at the beginning of an election cycle in which the Republicans ultimately won the national House vote by 51-45%, and gained 63 seats in the House of Representatives.

Rasmussen & YouGov/PoliMetrix tracked the House ballot weekly. Their numbers in the chart above represent monthly averages. Data compiled from Huffington Post Pollster, Real Clear Politics, and Polling Report.

In other words, Republicans performed a net 12 points better than the average of polling at this point in 2009 predicted. Furthermore, at this point in 2009, Republicans led Democrats in only 4 of the 38 generic ballot polls taken since the November 2008 election, as compared to leading in all but 2 of the 39 surveys conducted in the final month of the 2010 election.

Going back further in history, we see that early generic House ballot polling again badly missed the mark. Consider the 2002 Midterm, the first national election following the September 11, 2001 terrorist attacks, and another strong Republican year (they carried the national House vote 50-45%, and gained 8 seats, bucking the "President's party does poorly in midterms" trend):

Data compiled from Polling Report.

Republicans performed a net 8 points better on election day than early 2002 Midterm Polling suggested.  And like we saw in 2010, Republicans only led in 2 of 14 surveys conducted between the 2000 Presidential election and August 2001, despite leading in the overwhelming majority of polls taken in the final month before the election.

While early 2006 House ballot polling also missed the mark, it was more accurate than both the 2010 and 2002 early generic ballot surveys.  Democrats led Republicans in all 22 surveys conducted between the 2004 presidential election and August 2005, and by an average margin of 44-39%.

Data compiled from Polling Report.

But on actual election day, Democrats performed a net three points better than early House ballot surveys suggested, winning the popular House vote 52-44%, while winning 31 seats.

All of this suggests what we already know - polling conducted months, or even years out from an actual election, while helpful for understanding the current state of the race, shouldn't be counted on to reflect election day sentiments. The last three midterm elections all stand as evidence (to differing degrees) of that fact.