• Do Political Polls Really Matter?


    How many times, even on this newscast, have you heard the phrase:  according to this poll or that poll, a certain candidate leads the race for President by 3 points.... With a 4 percent margin of error.   Does any of that mean anything to you?

    How accurate are those polls?  And, who is answering these questions?

    There are, by my estimation, three billion political polls out there right now, all of them with varying results that stretch across the political spectrum no matter the race or the candidates.  There's even a book entitled "How to Lie with Statistics" written in 1954.

    People naturally tend to pay the most attention to the polls which best reflect their candidate's best chances.  In fact, both the Obama and Romney Campaigns have complained about inaccurate polling data since before the Conventions.

    The candidates aren't the only ones complaining.  We did an informal and very unscientific poll ourselves on my myfoxmemphis.com.  We simply asked, do you trust political polls?  The margin of error, well it didn't matter because we only asked the question with no parameters or rules.  Out of almost 600 people who responded, 93 percent of them said no.  They do not trust the political polling process.

    Here's what some of you told me on Facebook:            

    Fran Morgan Yocca - "no, the poll are useless unless you just want to make people believe their vote won't count."

    TbyrdMmira -  "anyone can find a favorable poll to fit their agenda so no, I don't trust them. I think the news media polls are a way to manipulate the sheeple."

    Jane Owens Brignole - "I do not trust, nor am i interested in polls. As a matter of fact I've never seen anyone being polled around my precincts."

    Berge Yacoubian of Yacoubian Marketing Research in Memphis understands the publics reluctance to believe what they are seeing, but explains that the process is indeed very scientific.  He, like other pollsters, has established guidelines that ensure the accuracy, within a margin of error of course, of their polls.

    Here are the Yacoubian guidelines for a good poll:

    Document #1

    Document #2

    Next Up: