Over the weekend, Obama campaign manager Jim Messina told the media that national polls don’t matter. According to Messina, the campaign is doing well in the important battleground states, and so national polls are less relevant than many would otherwise believe.

While this is not a bad point, of course it is simply partisan campaign spin to say national polls mean nothing. The significance of polls was shown last week, when National Review Online reporter Katrina Trinko wrote about how many national polls are biased…and that bias could mean the difference between Obama or Romney winning in November. According to Trinko, the major reasons for the sometimes obvious bias in polling could be due to several factors:

  1. Party identification shifts that will take place between now and November 6 may not be accounted for, simply because the pollsters don’t know how many people will switch.
  2. Question bias could be major factor – as Trinko writes, if a series of questions has hammered Republicans and made them look bad, a last question asking for party identification may be given inaccurately.
  3. Many Democrats are Republicans by voting style – for example, in Oklahoma, party ID is vastly in favor of Democrats, but it’s one of the most Republican states in the nation.
  4. Asking those taking the survey if they are a Republican or a Democrat instead of Republican, Democrat, or independents could lead to skewed results.

Trinko’s piece is important because grassroots activists need to not become discouraged when seeing biased polling. The one poll that matters is the one that takes place on November 6, and until that day is over no one can say for certain whether supporters of free markets and smaller government will win the day.