As I write (about 11:45PM Eastern) the Detroit News is providing an interesting example of what is either the limitation of exit polling or the limitation of polling on controversial race issues.
This article reports that, based on exit polling, “Proposition 2 [MCRI] is “too close to call.”
Voters who wanted to preserve racial, gender and ethnic preferences at public colleges and governments prevailed by about 4 percentage points, according to the poll sponsored in part by The Detroit News, WXYZ (Channel 7) and WWJ (AM (950). That is within the margin of error of the exit polling.
Curiously, in a later version of the same story by the same authors a bit later, the margin of error of the same poll was reported as 3%.
At the same time, however, the front page of the Detroit News has a box with running vote totals that, with 2741 of 5681 precincts having reported, MCRI is winning 58% to 42%.
There is still time for the exit polls to be proved accurate as more returns come it, but at the moment there would seem to be a disjunction between what voters told exit pollers and how they actually voted. Or maybe not so big, since the first Detroit News article on the exit poll reported above neglected to mention what the second article revealed: MCRI winning by 55% to 45%. [That seems like 10% to me, not “about 4 percentage points,” and greater than a margin of error of either 3% or 4%.]
There is also an interesting demographic breakdown of the exit poll results. Men supported MCRI 55% to 44%; women opposed it 58% to 39%. Whites supported it by about the same margins as mend did: 56% to 42%; blacks opposed it 86% to 12%. Republicans supported it 76% to 22%; Democrats opposed it 75% to 23%.
UPDATES
2908 OF 5681 precincts reporting (12:10AM): 58% to 42%
3818 precincts reporting (12:33AM): 59% to 41%, and with this the Detroit News has declared MCRI approved!
Now go figure: at 12:50AM Eastern the Detroit News home page contains a link to this article (although the time stamp on the home page says 1:50AM), whose headline announces “Both Sides Remain Hopeful in Affirmative Action Vote” — even though the paper’s own box score, on its home page, has declared MCRI the winner! In the article:
Both sides in the controversial move to ban affirmative action remained hopeful of victory late Tuesday night although exit polls showed the results too close to call.
….
Exit polls in the hotly contested race were too close to call. Early vote returns, with 34 percent of precincts reporting, showed that Prop 2’s passage was winning by a margin of 58 percent to 42 percent.
Oops. Now, about five minutes later, the article has been rewritten again (same URL as just before), and it now leads with the declaration:
The controversial proposal to ban affirmative action at public colleges and governments was apparently approved by Michigan voters Tuesday.
The Associated Press estimated that the ban, called Proposal 2, would win when all the votes were counted.
Finally!
The first quote above contain begins with the clear declaration, “Voters who wanted to preserve racial, gender and ethnic preferences at public colleges and governments…”
This seems almost like JSR himself could have written it.
Of course, the later stories all seem to focus more on “the controversial move to ban affirmative action.”
The editors must have come back from getting coffee.