« Hands Off: Some Have Concerns About Guv's Plan To Amend Constitution | Main | Your Indiana Hooooooosiers: Top Ten In B-Ball And A Bowl Game, Too? »

Rewriting History: Amos Brown Explains Indy Star's Mayoral Mistake

FlipflopRadio host and the Indianapolis Recorder columnist Amos Brown did some digging yesterday and discovered the reason why the Star had to revise its polling data for the local mayoral race. Here's an excerpt from his explanation:

"In e-mails I questioned the accuracy of the poll with Dennis Ryerson of the Star and Jim Tellus Channel 13 news director. They responded that they "weighted" the poll to reflect Black voter participation in Marion County at 16.5%.

"There is no data available on how Blacks vote or in what strength in Marion County. Given that roughly 35% of the Black population lives in a "white-majority precinct" its virtually impossible to track Black voting habits in their county, without doing sophistical exit polling.

"Anyway, I shared data from the Census Bureau on estimates of voting, by race, for the state of Indiana in the November 2004 election. They evidently consulted their pollster, Ann Selzer of Selzer & Co.

"Inexplicably, Selzer was using Census voting data from 2000 for Indiana. Data which was based on outdated 1990 census projections, not the final 2000 Census head count.

"I started calling this problem to the Star's attention around 3:20pm. By Channel 13's 6pm newscast, the poll had been recalculated, reweighed assuming that 20.5% of all voters will be Black. (An unrealistically low estimate).

"I have never known a media organization to publish a poll then pull and recalculate the numbers within hours after its release. I seriously question their assumptions of Black voting habits for this city/county. This is a close election and this is a sloppy poll. The Star and Channel 13 did a disservice to both candidates and to the Black community."

To hear more about the poll and the statistical mistake, you can tune into Brown's show today from 1-3 p.m. on WTLC (1310-AM).


Novice here but . . .

"Anyway, I shared data from the Census Bureau on estimates of voting, by race, for the state of Indiana in the November 2004 election. "

Why not use ( If available ? ) the "Census Bureau on estimates of voting, by race, for the state of Indiana in the November 2003 election.

Comparing a presidential election year vote totals to an off year mayoral total is like apples and oranges ?

Are the % of black voters who vote in a presidential election year the SAME as the % who vote in the off year election ?

In 2004 Marion county voted like this

Marion Dem - 162249 / R - 156072
Total - 318321 = 53% turnout

But for the Indy 2003 Mayoral election the numbers were...

D - 92763 / R - 55354
Total = 148117 or about 25% turnout.
(IF you figure 600,000 eligible voters )

So the question is - do the same proportion of voters by race who vote in a Mayoral and a persidential election the same or are they statistically different ?

Just curious

Why is it ALWAYS about race for Amos???

And while he was off ranting to the Star, did he bother to ask about Hispanic voters and how they were counted? Or don't they matter?

Don't you think it's important for a public opinion poll to accurately represent the public being polled?

Jen...although I agree with you in principal..Amos does play the race card to much..........and while I am typing...I have nothing against Amos and respect him as a person.

I guess this is as good a time as any to highlight the obvious:

1. Why does the Star continue to use an Iowa-based polling firm? I know Ryerson was familiar with Selzer from when he worked in Des Moines, but we do have folks who know what they're doing here as well. And they might even understand how Indianapolis votes.
2. Why would the Star go cheap in this case, settling for a sample size that would produce a large margin of error and would produce nothing useful about Council races? This is what we call a false economy.

Voting data by race for Indiana and any other state and the nation are only available during Congressional election years.

Why is it when poll results are against the D's it's always a "statistical mistake"?

Just wondering.

Why is it ALWAYS about race for Amos???

Easy Answer:

if the sample of white voters had not been large enough do you not think tom john would have bitched??

I have zero tolerance for Amos.

But even a blind squirrel finds an acorn once in a while.

The 400 respondents is perfectly fine, Riley...enough to extrapilate many facts and trends.

But ONLY if the 400 accurately reflect those who have voted in the past and plan to vote in two weeks.

Therein lies the rub.

My guess is, the Selzer group was either cheaper, or a Gannett "find." Their methodology is sloppy, proven by their withdrawal and recalculation, which is, sports fans, unprecedented. I got their call, as I posted here, and the questioner read the questions so fast and incomplete that I had to ask her to slow down and repeat herself four times.

I think the race is tight, but not on a level the poll showed.

Amos got one right...the racial makeup of the polling pool is statistically very important.


"Peterson declined to answer questions related to the race at his only public appearance today.

He presented an Excellence in Education Award at Indianapolis Metropolitan High School, a charter school he created, and did not make himself available to questions from a reporter."

Cash, I'm going to have to disagree. What possible trend could you extrapolate from this poll? It doesn't even tell you, within the statistical level of error, who's leading the race for mayor. And with respect to the Council races, the only question asked referred to voting for "the" Republican or "the" Democrat, completely ignoring the at-large candidates. This is pretty much a worthless poll.

Riley, I didn't say it WASN'T a worthless poll. Just that the baseline sample group was flawed.

And the council preference question was asked only AFTER a long-winded question about Monroe Gray's ethics. Which wasn't as much a question as an indictment. It wasn't a candidate-preference question, but a party preference question. Which is silly on its face. That's a potential February question, not an Oct. 21 question.

I happen to think, anecdotally, that the mayor's lead has shrunk dramatically, but nowhere near the numbers reported in the poll.

I think The Star manufactured this poll to generate talk, and Ch. 13 was a willing unindicated co-conspirator, to generate ratings.

How else do you explain this---Ch. 13's Mary Milz reported tonight on the council results from the poll (which were, as Riley noted, largely worthless). Nowhere in the newspaper or on Ch. 13 did they disclose that Milz is married to Star Editor Dennis Ryearson.

Which is hugely unethical. In no other major media market that I've viewed, would that kind of relationship be tolerated, let alone undisclosed.

And don't forget, again: the polling organization released its results, and within an hour, retracted, recalculated and re-released. And like the little girl Julia saw but no one else saw in plain view of Tully--no one has asked why this little lapse in credibility is accepted as norm. Lazy media--we're all paying the price for it, and it's an epedemic, spread to the electronic media via these insane alliances.

If anyone can find a similar polling tactic/release anywhere in the US in the last decade, please let us know.

Let's not lose sight of the fact that the mayor's campaign has been lackluster, and his silence on the council's nonsense, will undoubtedly cost him. It already has.

But this poll was and is a sham. The real pulse of the voters may indeed be what the poll says--we'll never know, because they did such a lousy job, and the alliance that produced and now promotes the poll is a tad unholy. In the media credibility sense, not the God sense.

Memo to Ch. 13 assignment desk: Mary had ought to never report on a story within that Star/13 alliance. Ever.

"Memo to Ch. 13 assignment desk: Mary had ought to never report on a story within that Star/13 alliance. Ever."

Memo to Cash: I don't get the conflict of interest in Milz reporting on a Star/WTHR poll.

Are you afraid that she'll oversell the importance of the poll because of her personal association?

Obviously, the paper and TV station are heavily invested in this project. What else could an overzealous Milz do - "No, listen, this is REALLY significant stuff..."

A real conflict would have to involve a story where Milz is reporting on the Star itself (not a joint project) and has the opportunity make the paper look better through her writing or analysis.

Of all the potential problems in local news these days, you've seized on a peripheral concern and focused on a couple of people you don't like.

Your dislike for the Star - that looks like the real conflict of interest here.

Note to 9:14 - A whole hell of a lot of us dislike, read hate, the Star. Or should I say, lots of us hate Gannett. Their website gets worse everyday. I read a lot of online newspapers and none but Gannett have the shameless, in your face, advertising that the Star site has. Currently we have this ridiculous Sprint football player that wastes even more of my time while I wait for the page to load as it runs to the corner and pulls back the screen. Why? Who knows? If and when I want cutesy cartoons, I'll find a site with them. I go to a newspaper site for - guess what - NEWS! There's more local news in the IBJ! I bookmarked the local news and obits so I can bypass the Star's nonsense.

Use the Firefox free web browser -- it can block most pop-up ads and exuberant Flash displays. It makes the Star website tolerable for viewing.


My journalism background tells me, a news reporter should disclose any bias. You'll often see NBC, for example, when reporting on defense contractors like those owned by GE, disclose that NBC's parent company is GE. It's basic reportorial and news management ethics.

Clearly, the Star/13 poll had a flaw. Somuch so that it was pulled an hour after it was released--then recalculated and re-rereleased. Which, in political polling, is never done.

So, regardless of the results of that poll, wouldn't you say, in fairness, that there's at least some controversy surrounding the Star/13 poll? That is methodology has been brought into question? Perhaps the pollsters hahve sufficiently answered those questions, but there's a smell there.

So, the TV station has the newspaper editor's wife report on the poll?

And the lead is, the Council races "may be close or tipping to the Republicans" in Milsz's exact words. Never mind that the at-larges have always followed the mayoral candidate. Never mind that the poll asked the council question, ont about at-larges, but about "the council" generically, without specifics, and after sufficiently criticizing Monroe Gray's stupid leadership.

Neither the poll nor the Star's story indicate another salient fact: the 2003 actual district council vote totals were about 55-45 Republican. Or, extrapalated, about the same or close to the margin their "poll" found last week. Wowie! Zowie! 'Our poll didn't ask about head-on races, but we determined the district races were about the same as the current council.' Stop the freaking presses.

Under no circumstances should the TV station and newspaper assign reporters to cover this poll, whose objectivity can be so easily questioned. Especially when the poll's veracity is already questioned.

I have met and like both Dennis and Mary. But the news leadership in this city began a long slow news judgment decline many years ago, and it continues. On multiple subjects--not just this one.

Post a comment

If you have a TypeKey or TypePad account, please Sign In

Find Stuff


Buy Stuff