The Detroit Free Press has a new poll on MCRI.
A thoughtful reader has a problem with it, or perhaps with another poll showing different results, or perhaps with all polls:
On October 18, 2006, you [the Free Press] reported on a Detroit Free Press poll of 643 likely voters taken between October 8-11. It showed 41% favor passing Proposal 2 (MCRI) while 44% oppose it. The poll showed that 15% were undecided or had no opinion. That poll has a plus-minus margin of error of 3.9%.
http://www.freep.com/apps/pbcs.dll/article?AID=/20061018/NEWS06/610180301/1001/NEWS
On October 6, 2006, the Detroit News published an EPIC-MRA poll of 600 likely voters taken between October 3-4, just 5-7 days earlier. It showed 49% of those polled favored passing Proposal 2 while 33% opposed and 12% were either undecided or had no position. That poll had a margin of error of 4%.
http://www.detnews.com/apps/pbcs.dll/article?AID=/20061006/POLITICS01/610060354/1022
These two polls, taken just 5 to 7 days apart, show a variance of 8% in support for the Proposal, a variance of 11% in opposition but a variance of just 3% in the undecideds/no opinion.
Does not that seem odd?
An EPIC-MRA poll of 608 likely voters taken between October 10-12, just 1-2 days after the Free Press poll, showed 49% of those polled favored passing Proposal 2 and 1% leaning to favorable for a total “yes” of 50%. That poll showed 39% opposed and 2% leaning to no for a total “no” of 41%. 9% were reported as undecided. That poll had a plus-minus margin of error of 4%.
http://www.woodtv.com/Global/story.asp?S=5535470
So the Free Press poll you reported on today and the most recent EPIC-MRA poll, taken just 1-2 days apart, show a variance of 9% in support for the Proposal, a variance of 3% in opposition, and a variance of 6% in the undecideds/no opinion.
How do you account for that? I am no political scientist or polling expert but it would appear that somebody’s polling is flawed.
The truth is all the polls are badly flawed. There is a high disincentive for poll respondents to reply candidly to such an incendiary issue where the risk of being stigmatized is so high. The polling experience in Washington state had support for affirmative preference ban disappearing entirely before the election but the ban went on to pass comfortably at the election.
The reason for the polling error phenomenon was expressed well by Bill Ballenger, Michigan’s premier political expert, in a quote published in the CMU student newspaper today. Ballenger was asked to comment on a recent CMU student poll which showed 56% of CMU students polled claimed to be undecided or uninformed about Proposal 2:
Bill Ballenger, Griffin Endowed Chair for Political Science, said many of the students who claim not to know about the MCRI actually have formed an opinion.
“There’s a good chance a lot of them are lying,” Ballenger said. “The news media has been so overwhelmingly politically correct and so biased toward defeating the MCRI, people are basically cowed into saying they’re against it and a lot of people have stopped talking about it,” Ballenger said. “They don’t want to be labeled as racist for opposing (affirmative action).”
http://www.cm-life.com/vnews/display.v/ART/2006/10/18/4535ac2bf1158
