Tuesday, September 25, 2012

Polling and the 2012 General Election – Is there a Reliable Poll – most probably not – Fact: Polls Based on Four year old trends Cannot be Accurate

What happens when one adjusts the polling data based on 2008 trends with current voter identification? - 1980. image: screenshot of www.unskewedpolls.com

There’s a feeling of Déjà-Vu about this general election cycle, if one were awake or living in the year 1980 – the economy was in deep trouble, the President a proponent of “Big Government”, entitlements were up, interest rates were through the roof, the price of gas was astronomical, as was food, and to top it all off there was a crisis in the Middle East. The polls throughout the 1980 Presidential election were either tied or had the President in the lead – right up to the night before the election – the prediction: too close to call.

Of course, the news articles, and nightly newscasts referred to the GOP candidate, Ronald Reagan, as a “clown”, and negative articles and broadcasts appeared “stacked” against the Republican challenger to President Carter. From the Milwaukee Journal Sentinel, September 5, 1980 – via the Washington Post: “Reagan Campaign Battles Slip-ups” - the article goes on to cite growing concerns within the campaign, the gaffes the candidate was making - the perception – the campaign was in disarray.

Sound familiar?

After the dust settled, Gallup explained the Carter loss as a “Dramatic Vote Changes Given Carter in 76, 80”(Youngstown Vindicator, December 21, 1980). Simply put, Carter won the youth vote and the ethnic vote, but lost his edge on the balance – his support among traditional Democrat voters fell from 82% to 69%, and he lost in a landslide.

Therefore, one had, an extremely re-electable President, by all polls, a challenger whose campaign was in trouble in late October, the GOP candidate made gaffes so often by the media standards, he was not electable - and yet – he prevailed.

Fast-forward to 2012 – and one finds Romney’s news coverage, with very limited exception – negative. He’s made a lot of gaffes and the polls are tied up or worse, the President is leading in three key swing states! – Romney’s Campaign is in disarray!

Trust in media has slipped to unknown territory – with these same pollsters offering a tied scenario – which allows broadcast to manufacture an image that the Romney campaign is somehow deficient due to the polls. These polls that are using samples (those surveyed) that are based off election statistics from 2008. Therefore in simple terms, this tests the limits of simple math.

In 2008, the Democrats had a large share of registered or identified voters, there were fewer registered Republicans, and Independents leaned Democrat, they made up a fraction of the vote. These statistics have changed over the course of the past four years, the Democrats have lost their huge lead over voter registration and the Republican’s have increased their share of the electorate, as have the Independents to a greater extent.

Therefore for a polling firm to base 2012 projections on the 2008 model may be a tad misleading.

Which is the reasoning behind the new website www.unskewedpolls.com - a site that is similar in scope to the model used by Real Clear Politics, which blends all polls and uses the average as a result. With the exception, the polls are re-calculated using voter registration/party identification from 2012. When one changes the percentage of the sample, one finds that candidate Romney is leading President Obama by an average of 7 to 13 points, as suggested by the unskewedpolls.com.

Although, one must question the accuracy of changing the original polling data, even basing that data on accurate samples, to project an outcome by using bad statistical data in the first place!

That said, the models used, even though the accuracy might be questions, give candidate Mitt Romney a lead that is identical to the results of the 1980 election. (Refer to the Milwaukee Journal Sentinel Gallup article in paragraph 2)

Therefore, it would be wise to take any polls, at this point, with a grain of salt, unless, of course, the pollster shares their methodology, showing an accurate sample of the electorate based on the most recent voter patterns and identity. Therefore, the results will be known on the morning after the November election.

If the statistics hold for the electorate as it now stands, and those numbers go to the polls and vote, regardless of whether or not a candidate is “likable”, regardless of how “smart” a candidate is, or “how rich” – then one would suggest that history might just repeat itself. One final thought:

The logic in 1980 was that although Reagan might be “likeable” – Carter would win on intellect (The Modesto Bee, September 28, 1980). The reasoning, people would not vote their wallet. This was a reverse on the angle that President Obama is more “likable” than Romney, and yet, the Obama Campaign is counting on the – economy not counting. Same election, different decade.

No comments:

Amazon Picks

Massachusetts Conservative Feminist - Degrees of Moderation and Sanity Headline Animator

FEEDJIT Live Traffic Map

Contact Me:

Your Name
Your Email Address