Posted: September 14, 2012
I have never been a big fan of surveys. I learned at an early age that many questionnaires are designed to solicit a particular answer that either satisfies a cause or helps a politician running for office. On the other hand, I gained knowledge through doing some research that not all public opinion polls are bad.
Over the next six weeks, some of us may be asked if we have time to be a part of a survey. With the presidential election less than two months away, I think it's important that voters understand who is asking the questions and whether the results can be manipulated.
Independent polls, including those conducted by Gallop, the Pew Research Center and others involving television networks and newspapers, are generally non-partisan. However, I usually find a way to avoid taking a survey if it's being conducted by a political party.
I have found that the easiest way to get out of having to answer survey questions is if I tell the pollster that I'm a member of the media. For whatever reason, the media has such a negative connotation that we are not even fit to properly answer poll questions.
I used to get frustrated when I heard that a poll only interviewed about 1,000 people to obtain its results. I thought, "How could 1,000 individuals provide enough information to predict the outcome of a statewide election that would involve a few million votes?" Although I thought I was pretty good at basic arithmetic, I likely didn't pay attention much during the discussion about probability.
What I've found is that even a small sample of 800 to 1,000 "likely" voters is enough data to make an educated guess as to the outcome of a fairly large election. It's only when polls ask those who are "registered" voters that some of the conclusions tend to be less accurate. Apparently, asking whether someone is a registered voter does not mean that individual actually plans to cast a ballot.
Lastly, it's important to know how a survey was conducted to know whether the process is being done correctly. If a surveyor doesn't explain the methodology, then I automatically think that someone is hiding something. In addition, I always like to know the possible margin of error whenever survey results are being released to the public. I figure if there's a three or four percent possible swing one way or the other, then an election is too close to call.
Good luck to all who take the time to answer questions from a pollster this election season. And feel free to use the media as a way to get out of doing the survey if you think it's a waste of your time.