140715124640-online-polls-620xaThe Democratic Party of 2014 is starting to feel like the GOP of 2012.  The GOP of 2012 saw Romney leading in national polls, yet trailing in individual state races.  The party insisted Romney would win and instead he lost by 4% in the popular vote and over 100 electoral votes (332-206).  Today, the situation is reversed.  Two days from what Democrats see as a midterm shellacking they have turned to the GOP’s 2012 tactic, attacking the polls and arguing poll bias. There are two problems with this strategy.  First, there is just no evidence the polls are biased FOR Republicans or even AGAINST Democrats.  Second, even if the polls are biased it might not matter.

Let’s take these issues one at a time.  First, let’s look at the pro-GOP claim.  Nate Silver has written significantly on this topic.  In his first piece, he found their is no consistent bias in Senate polling since 1992.  From 1990-1994 the polls favored Democrats, 96-2000 they were pro GOP, 02-04 they were pro-Democratic, in 06 they overstated the GOP’s position, in 08 the Democrats and in 010 and 012 they overstated the GOP’s standing.  Democrats refuse to concede the point.  Rather, they point to Nate Cohn who has written an article on it here, and feel good articles for Democrats here and here.  The problem is that even though Cohn questions Silver’s analysis, he cannot say for certain whether the polls are biased against Democrats.  He can only say he thinks they are.

The second issue assumes there is bias in the polls towards Republicans.  If this assumption is accurate the question becomes how much bias?  The RCP average of polls show the GOP leads by 3.8% in CO, 1.8% in IA, 2.2% in GA, 2.6% in AK, 6.5% in KY, 4.6% in LA and 7.1% in AR.  So in the closest race, Democrats need to see a polling bias of at least 1% to win IA.  Next up they need 1.1% to be competitive in GA, 1.3% in AK and so on. In other words, Democrats need a polling bias of at least 1% to win any of the states they currently trail in.  The flip-side of the same coin is if their is polling bias in favor of Democrats they might lose North Carolina and New Hampshire and see smaller wins in states like Virginia, Minnesota and Oregon.

Seeing significant bias this year is of course an open question.  In Silver’s analysis, he found the GOP only had a polling bias of over 2% in their favor four times.  To hold the Senate, Democrats need at least a 2% polling error in their favor.  This would allow them to hold CO, IA, AK and take GA and stay at 50 seats (pending the outcome in LA).  But again, there simply is no way to know for sure whether the polls are off this much, let alone at all.

This has not stopped Democrats from claiming the polls are biased.  Democratic candidates and party operatives point to internal polls showing challengers and incumbents either narrowly behind or running neck and neck with their GOP counterparts.  But when campaigns have to cite internal polling numbers to raise their prospects it is bad news.  The one saving grace for Democrats might be the polls have not consistently moved against them at any one time this cycle.  Rather, the polls have slowly moved two steps away from them and one step towards them before the cycle repeats.  The latest batch of polling this Sunday showcased the trend.

The above analysis assumes a consistent polling bias nationwide.  But if the polls are biased in just a few states it could also have significant impacts.  I would be remiss if I did not acknowledge that GA and LA are likely to head to run-offs in December and January and Colorado is utilizing an all-mail balloting system for the first time.  Further, the fact an Independent is running in Kansas upsets the traditional paradigm between the two parties in the ruby red state.

Democrats can hope the polls are biased in a few states rather than the whole.  It might allow them to keep the Senate.  No history of this occurring exists however.  Sure, polls in individual states have been off but they have never been off enough in enough races to switch who controls the Senate (at least since 1990).

Add all these factors together and it is little surprise Democrats are arguing the polls are biased.  They desperately need them to be.  If the polls are spot on Democrats will still lose the Senate and multiple close races.  If the polls are biased towards Republicans, Democrats might still lose the Senate.  But, if the polls are biased towards Democrats it is Republicans who will benefit and sweep virtually ever close Senate race.  It may even trickle down to close House and gubernatorial races.  If so, the Democrats 2014 nightmare scenario will have come true.




Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s