Sunday, September 30, 2012

Obama & Romney: Are the polls skewed in Obama’s favour?

You’d think the results of so many polls would result in a fairly settled expectation of how many voters will vote for one of the two presidential candidates, but not so.

The Republicans believe the polls are skewed and do not reflect the reality on the ground. Therefore, the race is much closer between Obama and Romney than the media say it is. 

Some Republicans think the spread in Obama’s favour is around 2% today, while many journalists are talking about 5%+.

So who’s right and who’s wrong?

It seems to depend on what you consider the breakdown of voters is between Democrats, Republicans and Independents:

The basic argument that the polls are biased begins with the 2008 presidential election outcome. 

Obama won by capturing an electorate composed of 39% Dems, 32% Reps and 29% Independents; this differential is known as a D+7 spread/party affiliation or D+7 electorate. 
Barack Obama also defeated John McCain by 7 points in the popular vote. 

In contrast, G.W. Bush defeated J. Kerry in 2004 with an electorate containing 37% Dems, 37% Reps and 26% Inds. 

The 2008 election was unique in many respects… A spread of +7 was a rare event in presidential elections and the largest in decades; it followed eight difficult years, two wars, a financial and economic crisis.

The same conditions simply do not exist today. 

Yet numerous polls are not only using the party affiliations of 2008 as a basis for 2012 polling but building on that outcome. 

Specifically, the national partisan breakdown has shifted from D=41.4%, R=33.8% and I=24.7% in 2008 to D=34.7%, R=36% and I=29.3% in 2010, and is currently measured at D=33.3%, R=37.6% and I=29.2%. 

Clearly this is a growing trend toward self-identification as an independent or republican and a material decline in democrat identification… 

The CBS/New York Times battleground polls of last week for FL, OH and PA had some curious weighting in its internal statistics… 

Reasonable and objective poll evaluators would be intellectually dishonest if a little sunshine wasn't focused on this poll's result. 

Many polls have surfaced recently which have party weighting that are also premised on equally questionable internals (and turnout models) and invariably the MSM pounces on their results to bludgeon Romney and his adherents.

And Rasmussen shows a much closer race:

The Rasmussen Reports daily Presidential Tracking Poll for Sunday shows President Obama attracting support from 48% of voters nationwide, while Mitt Romney earns the vote from 46%. Three percent (3%) prefer some other candidate, and three percent (3%) are undecided.

We’ll know come November 6.

So keep your fingers crossed.

1 comment :

  1. http://fivethirtyeight.blogs.nytimes.com/2012/09/29/poll-averages-have-no-history-of-consistent-partisan-bias/

    ReplyDelete

Thank you for commenting; come again! Let us reason together ...

Random posts from my blog - please refresh page for more: