Research Business DAILY Report --- October 29, 2012

Welcome to your Research Business DAILY Report. We are sponsored by Socratic Technologies.
It feels good to be back. More on this subject at the conclusion of today's report.
Election polling, what else?, will dominate research news for the next 10 days.
Last week, Doss Struse, the President of YouGov America & Definitive Insights, forwarded an
insightful presidential polling analysis by YouGov President Peter Kellner.
He tried to simplify noticeable variances in state-by-state and national polls following
the first Obama-Romney debate.
Kellner's polling beliefs include taking account of respondents' political partisanship.
He asserts that party ID reduces the risk of a rogue poll with a demographically fine
sample that is politically skewed.
Another reason to incorporate Party ID? Many pollsters say it can change sharply in response
to specific events; YouGov believes it changes slowly
Being registered to vote matters as well, so YouGov's sample undergoes the unusual extra
step of checking its sample against each state's list of registered voters.
Kellner emphasizes polling the same people before and after the debate.
YouGov polled almost 33,000 voters in September and reinterviewed more than 25,000 of them
after the first debate.
Some polls reported an 8-point Romney first-debate bump; YouGov found a one-point narrowing of
President Obama's national lead.
93% of YouGov's pre-debate Obama supporters remained in his camp after the debate; 94%
of Romney supporters were still behind him.
Kellner claimed that adjusted ONLY for demographics and with separate pre- and post-debate samples,
Romney achieved a five point post-debate bump in 25 states.
What can we look for through election day? With 10% response rates using telephone polling,
even a 1% improvement in response rate from Romney or Obama supporters, attribute it to
enthusiasm or whatever, can have huge polling impact.
Kellner's final point was about IVR polls, which cannot legally call the one-third of
voters with mobile phones and no landline.
"They are people who are keenest on Obama. Robopolls tend to produce higher figures for
Romney," concluded Kellner.
In the end, YouGov's polling conclusion about the impact of the first two presidential debates?
Neither was a difference maker.
Perhaps the bottom bottom line was pre-ordained.
Weeks ago, I heard that historically speaking, presidential debates rarely move voter sentiments.
Maybe daily tracking polls serve no goog purpose and misleading, except for the Obama and Romney
campaigns. And they've got their own internal polling that are not being made public.
Gilda Radner in character on Saturday Night Live, may have had it right about this year's
debate polling many years ago: Never Mind
Finally, thank you so much to the hundreds of who emailed messages of condolence on October
18 after my father's death.
As you can see, I haven't shaved since his funeral, and until mid-November I won't.
It is part of how I am honoring my father's memory. I hope you will indulge me.
That's your Research Business DAILY Report, sponsored by Socratic Technologies.
Have a great research day, and we'll see you tomorrow.