Friday, July 17, 2009

Pollsters. Who is the best?

When election time rolls around, there are always plenty of polls spewing out all kinds of data. Unfortunately, we're never really sure who to believe and whose methodology doesn't measure up.

The most talked about and most recent research I've seen was by Fordham University when they determined Rasmussen did the best of all the major polls in the 2008 election. Unfortunately, they only took into account national polling data and ignored the much larger and more important data sample of how everyone fared when it came to state polls in battleground states.

So, in order to figure out exactly who did perform the best, I decided to do a little research myself. The formula, while far from perfect, worked this way:

I determined the Margin of Victory (MOV) in eleven of the most polled/swing states of the 2008 Presidential election cycle. Obama states are listed in blue, McCain in red.

Depressing, I know.

From there, I used and to find the most up to date polls closest to election day from eight of the top pollsters in the nation. Following this step I calculated their Margin of Error (MOE) versus the actual MOV. I never used data older than October 16th. Finally, I calculated the average MOE and placed them in order.

Ladies and gents, the results.

(click for larger view)
So what are the takeaways?
  • Color me surprised that Rasmussen did as poorly as it did.
  • Quinnipiac is #1 with an asterisk due to their low sample size. But, I kept them in there since they are big in Ohio.
  • And the biggest shocker, Zogby didn't suck.
A major problem with this analysis? It only has one actual datapoint to use to gauge the pollsters - that being election day. Obviously, no polls were taken that day so we do the best with what we've got, that being polls taken in the days previous when some minds were still being made up. In reality, we all know polling is far from an exact science, so in this analysis we do what we can with the information we have available.

So there ya have it. Now you have a slightly better idea about which polls to keep an eye out for and whom to use as toilet paper.

No comments:

Post a Comment

No profanity, keep it clean.

Note: Only a member of this blog may post a comment.