Perhaps the most bizarre part of my college experience to date is my newfound love for statistics. And as any statistically minded person will tell you, there is nothing like an election year to get the data compilin’. What has been the most fun to analyze for me this election cycle is the Ron Paul campaign. Ron Paul’s campaign has beautifully exposed how the major news networks are statistically terrible.

The networks love unscientific polls. They love asking us, the voters, what we think. But the fun thing about unscientific polls is that they don’t work and are total crap. This is where Ron Paul’s campaign comes in to save the day. Ron Paul is a libertarian long shot for the GOP presidential nomination. He knows it, but he wants his voice heard and a handful of Americans agree with his anti-mainstream rhetoric. So he continues on deeper into the race. In all the scientific polls, Paul is not doing so well. The NBC Wall Street Journal’s end of January poll had Paul pulling only 4% national amongst registered republican voters.

Luckily for Ron Paul, the networks insist on resorting to unscientific polls. Ron Paul supporters tend to be younger and much more technically savvy. They have been flooding every Network internet poll with their choice for presidential candidate, normally leaving Mr. Paul with a healthy 10 percentage point lead over his rivals.

Nowhere was this more evident than during the ABC network’s presidential debate where they teamed up with Facebook to have people express their views online and vote for whom they liked. ABC would then report these results live and therefore allow the public to be part of the debate. But thanks to Ron Paul’s devoted following, he dominated every question posed on Facebook. The result was that ABC had no material and Facebook missed out on what seemed like a clever way to expand its audience.

Perhaps even worse than the Ron Paul supporters subverting ABC’s ill-conceived voluntary polling is the mainstream media’s misuse of scientific polling. The worst offender is MSNBC’s use of the Real Clear Politics Poll. Realclearpolitics.com releases a poll, which they create by averaging all the other major polls together. On the surface, this sounds great. One would assume that since they are all wrong in different directions an average should be able to correct for some of this, giving a “real clear” result. The issue however is something all Intro Stats students learn. Each sample is simply a model. It is not inherently true. So the samples come with confidence intervals, aka sampling errors. These errors really say that there is about a 95% chance that the sample with its confidence interval includes the true proportion. Some polls only adjust for a 90% confidence interval. Some adjust for 97%. So when Real Clear Politics averages all of these things together and thereby pretending that they are all the same, they are really adding apples and oranges. It would be like trying to predict the weather more accurately by averaging the results from a meteorologist, the Farmer’s almanac, and a tribal soothsayer. They create a false sense that this makes it more correct creating an unrealistic truth. This can have serious consequences in discouraging voters and political donors.

So, as we muddle through the months leading to the election, let’s try to keep our heads above the fray and thank God for the Ron Paul campaign’s constant proof that the media has no clue what they are doing.