"Have you heard? Susie Cooper is running for High School class president -- and she's giving out free cupcakes in the hall!". A winning strategy for any campaign in a closed community is to give away the most free stuff in order to curry votes and support. The classic popularity contest plays out and the High School gets a president.
Last weeks' Iowa straw poll seemed to play out disturbingly similar to this High School dynamic. There was a small, self-selected body assembled to vote, but instead of cupcakes it was Randy Travis being given to voters (Michelle Bachmann's tent featured the artist).
Straw Polls Are Useless
Straw polls were born from the practice of holding a straw in the air to see which way the wind was blowing. The analogy makes sense, but the practice fails. Here's how the Iowa straw poll works: get a non-representative group of about 17,000 people at a fair, let them visit candidate tents and then vote once they've bought tickets to a fundraising dinner. The winner of the poll gets to be seen as a front runner, gets extra press coverage by the over 300 reporters at the event, all of which gives them an edge on fundraising.
Unfortunately, in practice this type of straw poll simply shows which candidate is willing to spend more on pleasing straw poll voters than their stance on issues. Or to extend the analogy further, it is like holding up a straw and seeing who can buy a bigger fan.
From a data analysis perspective this is the worst way to design a random sampling model to determine public opinion. Any result derived from a system like this should be immediately thrown out given that it isn't random, candidates options aren't presented equally, and money is exchanging hands between candidates and voters. From a political system perspective, Robert's Rules of Order puts it best by stating these types of straw polls are "meaningless and dilatory" (Read the rules).
Straw Polls are Dangerous
The problem of a straw poll is that by default the winner is given a significant advantage due to press coverage. As well as the public opinion places some value in how people voted in Iowa since they don't know the details of how people were swayed at the actual event. Social proof is a powerful force, we look to our friends or influential groups to determine everything from what we should read (NY Times Best Seller) to political candidates.
What Winning the Straw Poll Does
Even though Michele Bachmann won by 1%, it was enough to position her as a 'front-runner'.
"Rep. Michele Bachmann won the Iowa Straw Poll Saturday, affirming her status as a top-tier candidate in the Republican race to challenge President Obama in 2012." - Fox News Headline.
Iowa Straw Poll: 17,000 votes cast
Michele Bachmann 28%
Ron Paul 27%
Tim Pawlenty 13%
Rick Santorum 9%
Herman Cain 8%
Google Web Search
Based on millions of searches through Google.
Looking at web search data, Ron Paul beats Michelle Bachmann by 2 search points over the last 12 months, which translates into well over the 17,000 influenced voters that determined the Iowa straw poll results. It should be noted that Bachmann's search interest only spikes to a level equal to Ron Paul's after the straw poll results -- another benefit of the straw poll victory.
But even this data is misleading as we can't tell if these searches are being done by people who like or dislike the candidates.
The Safest Data
Candidates try to skew the numbers for their own benefit, straw polls, and search engine results, are not great measures of candidate's chances of winning. Lots of data will be thrown around in the months to come so it is important to remember that this is not a high school election. Be careful of what you see from every poll and pundit. Remember that the safest data is your own fact-based opinion.