Won't get fooled again, Part Deux

I have addressed the attempt to deligitimize the Obama election by John Ziegler through his project Howobamagotelected.com. Nate Silver has called him out and made him answer a number of questions. A couple of notes, first Ziegler is a talk radio personality and has a number of traits that work well in that medium, such as publicity through confrontation. While Silver is doing a good job of coronering Ziegler from an Objective position, Zieglers audience isn't necessarily moved by the objective. Secondly, Silver made an initial mistake in his haste to call out the Ziegler survey. He used the term as a push poll. That may not have been the most appropriate term. Ziegler is trying to use that faux pas as a way to undermine the other very objective criticisms Silver is making.

Ziegler is out with a second poll to "answer the left" although this may be a marketing way to keep the story fresh. The crosstabs of the new poll by Wilson Research is here http://www.w-r-s.com/blog/wp-content/uploads/2008/12/ziegler_news-coverage-survey_xtabs2_081125.pdf
from perusing their blog they appear to be a rightish anti media organization.

You can also see Silver and Ziegler debate on Breitbart.tv here

I have taken a few minutes to go over the crosstabs and a number of my previous suspicsions seem to hold up. http://mysteryj.blogspot.com/2008/11/wont-be-fooled-again.html

The biggest point is that the more recent a story is the greater the chance the voter got it correct. The best example of this were the questions pertaining to Sarah Palin all three questions were stories that broke within the last 6 - 7 weeks of the campaign. More interesting, and one won't see this promoted by Ziegler, There was great similarity between Obama voters and McCain voters.

Pregnant daughter Obama voters 94% McCain voters 94%
150K Clothes Obama voters 82% McCain voters 82%
Russia from House Obama voters 83% McCain voters 75%

All events became news relatively late in the cycle. Doesn't appear to matter who the voter voted for. Now note the 8 % difference between McCain and Obama voters. This appears to be abou the midrange level of difference when the question in unflattering to one side or the other.

McCain Houses Obama voters 77% McCAin voters 70%
McCain / Keating Obama voters 49% McCAin voters 42%

Two notes 1) see the 7% difference between Obama voters and McCain voters on these questions that are negative toward McCain. Similiar to the 8% the Palin question above.

More striking notice there is a 27% difference regardless of the voters on the two questions. The first question was an event that occurred nearly 20 years ago the other in mid August this year. The 70 - 77% correct rates are comparable to the response rates of the Palin questions all of which occurred in the same time frame.

Questions on Biden

Obama tested Obama voters 50% McCain 60%
Plaigerism Obama voters 27% McCain 33%

Again two notes questions negative toward Biden. McCain voters get it correct by 7 - 10 % more than Obama voters. Very consistent with questions regarding Palin and McCain in reverse. Also note the 23% and 27% difference in correct responses on the current question versus the older question.

The questions on Obama are a little bit more hazy. However there is one clear distinction. Regarding the question of "spreading the wealth"

Obama voters 81% McCain voters 87%

Again we are looking at a 6 point spread with the candidates supporters slightly underperforming. This statement came out in the last few weeks of the campaing. Questions re: Obama dating from earlier in the year and a decade past had correct responses much lower. With the exception the Bill Ayers question which has its own issues that can be addressed seperately.

The results of these two surveys may infact be interesting, but probably not what the author intended or will promote. First voters on these particular questions tend to do about 8% better on questions that frame the opposing candidate in a negative light. One can speculate that negative information is more readily remembered when it reinforces an image or narrative that is already accepted. I am not sure one can speculate much. The secondly clearly indicates that these voters both Obama and McCain had much better recall regarding incidents that are more current.

I am not certain either of these conclusions can be considered to earth shattering, interesting yes.