« Is GPS One More Way We Voluntarily Surrender our Privacy? | Home | Cold Calling for New Consulting Clients »

The Evil Art of Election Polling

By crooky | January 9, 2008

I am a man that believes in the utility of statistics. I know enough about them to both fear and respect them. Statistics are like guns - used in the right hands, they’re can be as precise as a surgical tool. In the wrong hands, they can twist and bend the truth to the interest of the people manipulating the information. With a looming US election, you’re going to hear a lot of polling results predicting that the outcome of the election one way or another. What you might not know is how this information is collected and what motivates people to answer.

There are three times of polling around elections: random sampling, entrace polls and exit polls. Each of these tools can be used for good or evil. Let’s talk about them individually.

1. Random Sampling

Newspapers, political parties and broadcasters engage the voting public well in advance of an election to give themselves some intelligence on the upcoming fight. These kinds of surveys ask a lot of questions around issues that are expected to show up in the campaign speeches and to more generally identify local issues that candidates will have to respond to as their take their campaign to the communities they wish to represent.

By looking at the numbers, pollsters can predict which jurisdictions are going to fight over particular kinds of issues. This gives political parties the ability to target their campaign speeches to regional interests. It also gives the news organizations an opportunity to predict what the debates will be about. There isn’t really much one can do to twist random sampling other than to ask leading questions such as:

Who do you plan on voting for? The Democrats with their pedophilic tendancies, or the Republican party with their baby-eating platform?

Data at the early stages showing a potential victory for one party or the other can effectively put the party that is predicted to lose on the defensive - taking them off their intended mandate and possibly leading to some mud-slinging against the other side. Voters don’t benefit when the rhetoric descends to the mud-slinging level.

2. Entrance Polls

Entrance polls essentially ask people how the are planning to vote as they go into a polling station. There are numerous problems with this approach since a lot of people believe very strongly that their voting preference is secret, before and after the fact. This means that the participation rate of entrance (and exit) polls is going to be lower than normal. This will throw off the reliability of the data. Similary, people may be motivated to lie about who they’re going to vote for to throw off pollsters.

In effect, the only purpose of the entrance poll is to give news reporters the ability to roughly predict the outcome of an election while the election is in progress. It’s worth noting that the margin of error on entrance polls is higher than for exit polls. I’ll let you judge how much stock to put into that data.

3. Exit Polls

Exit polls are a much different animal because, if people tell the truth when they respond - the exit polls can provide a pretty accurate record of how people actually voted. In Canada, that might not seem like a big deal but in developing countries or countries just transitioning into a democratic system, exit polls can be a valuable tool in the fight against corruption and election fruad. In fact, exit polling was an invaluable anti-corruption tool in the Venezuelan recall referendum (2004), the Ukrainian presidential election (2004), and the 2004 U.S. presidential election controversy.

Exit polls can also be mis-leading and report false reports, resulting in embarassing victory speeches by people who didn’t win (ahem, Al Gore). In essence, I think you can see that polling around elections can have a spectacular impact on the outcome of the election. There are some people that believe polling and media influence are a hindrance to the democratic process. There are others that believe polling and that kind of research are harmless.

I’ll leave the decision to you. However, if you don’t believe that polling is a good thing, you can always refuse to participate when they call or approach you. Personally, I never tell them. That’s private.

*********************
Aaron “Crooky” Cruikshank is the Principal and Founder of Friuch Consulting. He has written professionally about science and technology for ten years.

Share/Save/Bookmark

Topics: Research Methodologies |

Comments