Maintaining Optimism in the Face of Reality. Occasional observations on the state of the world, society, business and politics. Usually anchored by facts, always augmented by opinion.

Minnesota Polling Bias from the Strib?  | e-mail post

On a more local political note, I am a little disappointed to see our Republican leader calling for the resignation of the Star-Tribune pollster for chronically underestimating (misunderestimating?) Republican turnout. I can see publicizing the inaccuracy, but I think taking Rob Daves to task personally was intemperate. On the other hand, it probably makes for better copy, and thus gets better coverage. Unfortunate rabble-rousing for the sake of publicity.

We should all appreciate the importance of accurate polling in a swing state like Minnesota, given that it can influence the mood and momentum of each national campaign, as well as where advertising dollars will be spent and candidates will visit through the election. I can also appreciate that the Strib does seem to have a bias against Republicans. The Pioneer-Press account of the story helpfully mentions:
"An April poll by the Star Tribune showed Kerry with a 12-point lead in Minnesota. But a May poll by the Pioneer Press and Minnesota Public Radio gave Kerry a 3-point lead, and a poll in June and July for the 2004 Elections Project at the University of Minnesota showed Kerry with a 2-point lead."
Even the Star Tribune acknowledges:
"In [2002 in] Minnesota, the [Star-Tribune] Minnesota Poll had an average error rate of 3 percentage points on where a candidate would finish. Zogby International, which also did polling in Minnesota, had an average error rate of 4 percentage points [also seeming to skew Democratic this year]. Mason-Dixon Polling & Research Inc., which did polling for the St. Paul Pioneer Press and Minnesota Public Radio, had an average error rate of 2 percentage points. Only Mason-Dixon's final poll correctly foreshadowed Norm Coleman's victory over Walter Mondale in an abbreviated race that remained fluid in its final days.

The final Minnesota Poll, conducted on the Sunday and Monday before the election and published with the election results on Wednesday, showed a 45-45 percent tie.

Fifty-four percent of the other states' polls had an error rate of 2 percentage points or less, putting Zogby and the Minnesota Poll below the national median in accuracy, according to the council."
I am guessing they may do results weighting based on an outdated weighting, or possibly ignoring the apparent Republican bias among independents. See comprehensive Minnesota polling results at RealClearPolitics.com (RCP is a great source if you want to do your own handicapping of the electoral college race.). Another interesting read is this analysis of consistent biases in national polls. The author demonstrates that Time, and unsurprisingly, Fox, have a systemic pro-Bush bias while Zogby and the Economist have an even larger systemic pro-Kerry bias.

I'm just looking forward to the next Mason-Dixon poll of Minnesota, as they really do seem to deliver the goods here in Minnesota, and seem to be the most accurate of any national firm.

e-mail post | Link Cosmos | [Permalink]  |  | Sunday, September 12, 2004
Comments: Post a Comment

This page is powered by Blogger. Isn't yours?