Election 2008 Polling Accuracy

July 13, 2007 by  
Filed under Opinion

Watching the change in the winds this week in regards to the Republican race this week was “anti-climatic”. We have all heard the “one time front runner McCain” stories all over the news this week. This morning on NPR I heard an interesting piece on the John McCain topic. I caller from Michigan had called in to give his two cents to the debacle, and then questioned the host and panel about Ron Paul coverage. The host’s response is as follows:

“Laugh” “Sure, Dr. Paul is leading in the Internet polls, but in the “scientific polls” he isn’t a blip.”

He then changed the subject and continued on.

My point today is to call in the accuracy of these polls and how they are used to persuade the nation. Lets look at the stats of some of these polls.

  • American Research Group
    • 600 likely Republican primary voters (those saying they will definitely vote in a primary or participate in a caucus in 2008) conducted July 9-12, 2007. The theoretical margin of error for each sample is plus or minus 4 percentage points, 95% of the time.
  • USA Today Gallup Poll

    • 2nd Link
    • This was a phone survey (I would assume land-line) and a margin of error of +/-5%.
    • 394 people who “lean” to vote Republican.

We need to find a method of better judging the candidates we have running. Everyone has heard the argument that Ron Paul votes are getting left out because his supporters are “the mobile age” and don’t have the traditional land lines. Another argument is that Internet polls can’t be trusted because you don’t know if they are registered voters (oh, and supposedly Internet users are not registered voters…I know I am).

A database already exists of “registered voters”. How legitimate would it be to create a website that took your “registered id” and required you to log in? What is your opinion?

Be Sociable, Share!