As the Virginia gubernatorial race rapidly approaches a series of new polls have come out showing divergent results. The most recent survey from Quinnipiac showed Terry McCauliffe leading Ken Cuccinelli (R) 43%-38% among registered voters. Surveys taken in late April by the Washington Post and NBC/Marist among likely voters showed Cuccinelli ahead. So what explains this difference? I have a few thoughts on this.
First-off, it is a little early for many surveyors to be using likely voter models to predict turnout for the race. Yet, both the NBC and Washington Post surveys are using them. Since off-year constitutional elections tend to have lower turnout than other elections it is no surprise the survey is showing more devoted conservative Republican voters participating. According to the Washington Post’s internal numbers 28% of likely voters identified as Democrats, 26% as Republicans, about 30% as Independent and the rest as other or did not answer. Contrast this with Quinnipiac’s survey which looked at a larger sample of 1,200+ registered voters. Among those sampled well over 30% identified as Democrat, only about 25% as Republican and the rest as Independent or other. So Quinnipiac’s survey leans more to the left by default. This is a national trend as many occasional voters register as Democrats and thus tend to skew registered voter samples compared to the actual electorate.
Another factor is where the Independents lean in each survey. In the Quinnipiac survey Cuccinelli and McCauliffe attract about 35% of Independent support. In essence this creates a net draw among key swing voters and base turnout determines the race. If this is true than McCauliffe has an edge simply by the fact there are more Democrats in the electorate than Republicans. The NBC survey also showed the candidates deadlocked among Independents and thus base turnout is giving Cuccinelli the edge in the survey. In the Washington Post survey which found Cuccinelli up five among probably likely voters and 10 among those most likely to vote Cuccinelli has a wide lead. Independents seem to be leading the difference in polling results.
Lastly, ideology is playing a role in each survey. In the Washington Post survey a majority of respondents said the Federal Government was doing too much and conservatives made up about 35% of the electorate. In the Quinnipiac and NBC surveys those numbers were very different with fewer voters identifying as conservative or believe the Federal Government is doing too much. Voters who are conservative and believe the Federal Government is doing too much fueled GOP victories in 2009 and 2010 and could help fuel a Republican victory in Virginia in 2013.
There could very well be other reasons why the survey results are diverging. It could simply be random chance or the way the pollsters are weighting the results. Regardless, survey results are starting to diverge and that is sure to fuel debate well into the future about who is going to win the Virginia race but more importantly, how polling should be handled in an ever-changing political and demographic environment.