Friday, October 19, 2012

The Problem with Polling and the Beauty of the Free Market

Yesterday Gallup released the latest poll related to the Presidential election happening in a matter of a few weeks. One of, if not the, most trusted sources of polling data and information has the Republican challenger Mitt Romney ahead in the polls by %6-7. Even though they had recently changed their polling methods to be more accurate, there are naturally many who reject the initial implication of this result or argue to Gallup's inherent neutrality.

As a policy wonk it is usually frustrating to see the media swarm like piranha on new polls that get released, either ripping it to shreds or worshiping it as the end-all be-all source of pundit nourishment.

First, we all need to calm down about these results. They are a representation of only a portion of the registered voters and, like any poll, is an estimant of current sentiment. No poll is perfect, it is guess work of taking a "statistically representative sample of a population" and everything can go wrong from biased phrasing, to poor computer mocks to sampling error.

It really all comes down to statistics and statistics have patterns and can be skewed. On one hand, as noted by the New York Times' 538 blog, the poll results provided by Gallup are "deeply inconsistent with the results that other polling firms are showing in the presidential race, and the Gallup poll has a history of performing very poorly when that is the case."

On the other hand, other predictive models with higher performance numbers as indicators are swinging the other way. This includes everything from other polls, the illustrious "7-11 cup poll," and the prediction markets.

Prediction Markets can easily be described as an online gambling forum where people bet on the likelihood of outcomes. People come to these websites, mainly in other countries since there are legal issues in the United States, and bet on events creating "market prices". Like commodities markets, the current market prices can then be interpreted as predictions of the probability of the event or the expected value of the parameter.

One such prediction market is a site known as Intrade and they explain how it works on their site;
Stock exchanges find the price of stocks, and futures markets find the price of commodities. Prediction markets find the probability of something happening - a predefined, uncertain future event. Will the financial markets be up today? Will a certain candidate win the next election? Who will win the Academy Awards? If you have an opinion on what will happen then you can make a prediction on Intrade. Predict correctly and you can win real money profits.

Say what you will about the various methods of getting a pulse of an election but when people have 100s to 10000s of dollars riding on the election, they do not take it frivolously and, traditionally, have been amazingly accurate in their predictions and usually are not tapered by the biases of pundits and talking heads - no one has a boss or a bottom line that effects their view of the election. Intrade currently ranks Obama at 61.3% to Romney's 38.6%, or did at the time of this post. Did I mention that these numbers change daily as a result of news, new polls and economic indicators?

There is a great article in the New York Times about these types of markets used as predictors. as argued by Justin Wolfers, a University of Pennsylvania economics professor with anecdotal evidence.
The reason that I follow prediction markets on election day, because they are typically way ahead of the talking heads. For instance, I was live-blogging the 2008 election count for The Wall Street Journal, and called the election at 9:49 p.m. — based on prediction market data. The networks didn’t call it until 11 p.m.
Polls, cups and markets as indicators of the future will never be 100% accurate and the real point here is that no one should look at any of these indicators as "gospel." So let's all relax and move forward to the big day.

No comments:

Post a Comment