Polling Under The Microscope

[LS]

Campaign Puts Polling Methods Under Microscope
By Alan Elsner

WASHINGTON (Reuters) - - One poll last week had President Bush leading Democrat John Kerry by 13 percentage points while another showed the candidates tied in the U.S election race. What is going on with public opinion polls?

As U.S. voters are swamped by a blizzard of rival surveys in the final weeks of the Nov. 2 presidential election campaign, pollsters themselves are facing increasing questions about their methodology and accuracy.

Polls do not only measure campaigns. They may also affect them. If one candidate starts trailing in the polls, some of his supporters may become demoralized and decide not to vote.

"Polls impact how voters feel about a candidate, especially because of the huge coverage they attract on cable TV and the Internet," said Dean Spiliotes, director of the New Hampshire Institute of Politics at Saint Anselm College in Manchester.

"In many respects, they have become a campaign tool for mobilizing a candidate's base and demoralizing the opposition," he said.

Moveon.org, an Internet-based group working to defeat Bush, took out a full-page advertisement in Tuesday's New York Times attacking the Gallup poll that put Bush ahead for employing faulty methodology that overrepresented Republicans and undercounted Democrats.

Some say the gyrations in poll results reflects a volatile and undecided electorate.

"Voter opinion has become quite unsettled in the past month. The number of swing voters has actually grown from August to September, which is not what we usually see," said Andrew Kohut who directs polling for the Pew Research Center.

Pollsters acknowledge that there are many ways errors can slip into their surveys including sampling errors, interview bias and poor questions. Even the order in which questions are asked and slight variations in wording can produce vastly different results.

"Some of us try to hide behind the concept of a statistical margin of error but I don't use that phrase because the errors are in fact unquantifiable. There is no fixed, finite, measurable margin of error," said Humphrey Taylor of Harris Interactive, a New York polling company.

PART SCIENCE, PART ART

John Zogby, who will begin polling on the presidential race for Reuters early next month, said: "For me, polling is 80 percent science and 20 percent art."

Pollsters are each in their own way trying to come up with a cross-section of the electorate that will actually show up to vote on Election Day. But each company has a different method of deciding who should be counted as a "likely voter" in an election in which probably only around 50 percent of the electorate will vote.

For example, pollsters ask voters a number of screening questions aimed at excluding respondents unlikely to vote. These may include: did they vote in the last election, do they remember who they voted for, do they know where people in their neighborhood go to vote, are they registered with a particular party and if so which?

Some companies simply exclude anyone failing to respond appropriately to one or more of these questions. Others, including Gallup, statistically weight participants in the poll depending on their responses to these questions.

"We either screen people in or out of the poll but we don't weight their responses. Empirically, we've found it produces better results," said Taylor.

The other major issue dividing pollsters is whether they should include weighting for party identification -- that is whether they should statistically adjust their raw numbers to reflect preset assumptions about the numbers of Republicans, Democrats and independents who will take part in the election.

All pollsters weight their samples by gender, age, race and other factors such as income, geographical location and education level to produce what they believe is a true reflection of the electorate.

Zogby is a staunch advocate of including party identification in his statistical method.

"I know that to some pollsters I am a heretic but I have found that weighting for party ID is a proven way of ensuring you have a proper sample," he said.

Other pollsters criticize this, since they say that people respond differently when asked which party they belong to depending on events in the campaign.

[LS]

Here's a perfect example...this is a question and the 3 answers that were given to choose from that I ran across on an online poll:

Should U.S. troops get involved in the upcoming Iraqi elections?

*They shouldn't leave until Iraq is at peace.

*No, it's too dangerous.

*It's too late to leave now.

Ummm...exactly in what way do any of the possible answers pertain to the actual question??

And this should further blow polling results out of the water :

New Voters Flooding Election Offices

By ROBERT TANNER
AP National Writer

New voters are flooding local election offices with paperwork, registering in significantly higher numbers than four years ago as attention to the presidential election runs high and an array of activist groups recruit would-be voters who could prove critical come Nov. 2.

Cleveland has seen nearly twice as many new voters register so far as compared with 2000; Philadelphia is having its biggest boom in new voters in 20 years; and counties are bringing in temporary workers and employees from other agencies to help process all the new registration forms.

Nationwide figures aren't yet available, but anecdotal evidence shows an upswing in many places, often urban but some rural. Some wonder whether the new voters - some of whom sign up at the insistence of workers paid by get-out-the-vote organizations - will actually make it to the polls on Election Day, but few dispute the registration boom.

``We're swamped,'' said Bob Lee, who oversees voter registration in Philadelphia. ``It seems like everybody and their little group is out there trying to register people.''

Some examples, from interviews with state and county officials across the country:

New registered voters in Miami-Dade County, a crucial Florida county in 2000, grew by 65 percent through mid-September, compared with 2000.

New registered voters jumped nearly 150 percent in Cuyahoga County (Cleveland) in Ohio, one of the most hard-fought states this year.

And that's with weeks left until registration deadlines fall, beginning in October.

Curtis Gans at the Committee for the Study of the American Electorate said a clear national picture won't emerge until more applications are processed next month. And Kay Maxwell of the League of Women Voters cautioned that some years that promise a boom in new voters turn out to be duds on Election Day. The danger is that new voters may not be as committed to showing up at the polls as longtime voters.

``Turning people out to vote is tougher than getting them to register,'' said Doug Lewis, who works with local election officials as head of The Election Center, a nonprofit group.

Rural areas, which trend conservative and Republican, aren't necessarily reporting the same growth as urban, more liberal and Democratic strongholds: Brazos County, Texas, hasn't beaten its 2000 numbers so far, though officials said applications are now rolling in. The state of Oklahoma, however, saw new registrations in July and August increase by 60 percent compared with four years ago.

Oklahoma officials said they had 16,000 new Republican registrations, 15,000 new Democrats and 3,500 new independents. In Oregon, where new registrations grew by 4 percent from January through Sept. 1, Democrats outregistered Republicans two-to-one.

Lewis and others say that no matter what the partisan breakdown, the registration boom is real - driven by a swarm of organizations such as Smack Down Your Vote (a professional wrestling-connected campaign), Hip-Hop Team Vote, traditional groups like the League of Women Voters; party-aligned groups such as America Coming Together, made up of deep-pocketed Democrats; and many, many more.

``There seem to be hundreds of them,'' Maxwell said.

The groups' focus is on states where the vote was close in 2000, but even in several states where the election isn't as competitive, officials say they are seeing new voters register in higher numbers. Officials in El Paso County, Texas, Maryland's Montgomery County, a suburb of Washington, D.C., and California's Los Angeles County said registration numbers are on pace to be higher than 2000.

In many jurisdictions, administrators complain that the crush of new registrations is overloading staff.

Clerks have hired extra workers in West Virginia, Ohio and Colorado. Philadelphia borrowed employees from other city agencies and started working overtime two months earlier than the usual post-Labor Day push.

In Greenbrier County, W.Va., deputy clerk Gail White said she's never seen so many people register in her 10 years working elections, and despite extra staff she's still behind on processing new and absentee voters. ``I get them all typed up, and the next thing I know, here comes another pile,'' she said.

The reasons seem clear - groups on all sides were energized by the close election of 2000, which proved to doubters that a handful of votes can swing an election. In 2000, 9 percent of voters, roughly 9.5 million people, said that was their first time casting a ballot, according to AP exit polls.

``It's the high-growth areas, the suburban and exurban areas in those battleground states,'' said Scott Stanzel of the Bush-Cheney campaign. ``There are opportunities there because there are so many new residents to register.''

The GOP has launched a volunteer, precinct-by-precinct effort in swing states, with separate help from a Republican-aligned group, the Progress for America Voter Fund.

Democrats, who've consistently made turnout efforts the foundation of their campaigns, are devoting huge amounts of resources, too. America Coming Together focuses solely on registering and turning out voters.

The McCain-Feingold campaign finance reform law has boosted efforts, too. It cut off unlimited ``soft'' money to the parties, diverting some of that cash to community-based groups.

In Missouri, the result is that what used to be a mostly volunteer-driven voter-registration effort by the Missouri Citizen Education Fund has blossomed into a bigger, paid-staff operation, said executive director John Hickey. Funds jumped from a few thousand dollars a year to $250,000.

Focused on poor, black neighborhoods in St. Louis, mid-Missouri and rural areas, his staff went from registering a few thousand new voters in 2000 to at least 50,000 so far this year, Hickey said. In 2000, George W. Bush won the state by less than 80,000 votes.

[Roland]

"Polls do not only measure campaigns. They may also affect them. If one candidate starts trailing in the polls, some of his supporters may become demoralized and decide not to vote."

That statement has always been my feeling on them, that they're used as a manipulative tool. The fact that there's a screening process at all is the primary tip off. The sample question you posted is another good example of how questions are weighted to purposely slant towards whichever special interest party initiated and is funding the poll.

I have to ask who's funding Gallup of late. First there was the case where after the RNC, they reported Bush with a 13 point lead when every other poll either showed Bush with a minimal lead or in a virtual tie with Kerry. Today they released their poll on the debate. They show that Kerry won overall on presentation, but on all the other points such as "believability" and who was more factually correct, they show Bush won by large percentages on all those points the complete opposite of every other poll I've seen.

[Roughneck]

That just stands against all reasoning. Saw the same deal with the Newsweek poll. Again, overall Kerry won by a large margin, but on nearly all individual aspects, Bush won That's like saying that a football team scores more points in all four quarters but the other team still wins.

[LS]

Roughneck said:

That just stands against all reasoning. Saw the same deal with the Newsweek poll. Again, overall Kerry won by a large margin, but on nearly all individual aspects, Bush won That's like saying that a football team scores more points in all four quarters but the other team still wins.

And that's the thing about polls that makes them highly questionable...
Like one of the questions asked was- who's the more consistent?? Most people probably would answer the DuHbya- because he is- even after he's proven wrong beyond a reasonable doubt time and again...he always keeps insisting he's right (and made no mistakes ). But the thing is- notice there's no follow-up question asking people if they view that as a positive or a negative- it's just left at and the way I see it is that leaves it very much open to interpretation. I mean I'd have to say the DuHbya's consistent...consistently wrong- so despite the fact I might answer he's consistent, in my viewpoint that doesn't necessarily make being 'consistent' a good thing.

Polls are one of the more amusing aspects we have...the questions are either purposely screened and weighted to slant results in a certain direction, the questions are so poorly worded that people simply choose 'the closest' answer to what their viewpoint actually is or the questions are so general (such as the above question) that they falsely present 'black or white' readings- and things just aren't that simple in a world chock full of 'gray.'

[Roland]

Good points all. Arianna brought up a few more good ones today.

Media's Worship Of Polls Is A Farce
Arianna Huffington
October 12, 2004

It's time to pull the plug on the media's treating public opinion poll results as if Moses had just brought them down from the mountaintop.

Over the last month, media coverage of the presidential race has been driven by wildly vacillating poll numbers. For example, Newsweek has John Kerry going from 11 points down in its Sept. 4 poll to 2 points up in last week's poll, while Gallup went from Kerry's trailing by 14 points on Sept. 16 to dead even on Oct. 4.

Yet at the same time that Gallup had President Bush 14 points ahead, the Pew Center poll had the race all tied up. And when Gallup had Kerry pulling even, Pew had the President holding a 7-point advantage.

But no one in the media says, "Wait a minute! Both of you can't be right." They just dutifully report the latest numbers and set out to explain what they "mean" without any attempt to account for the huge disparities.

The dirty little secret of the polling industry is that all too often its findings are based on flawed methodology and dubious assumptions. Take that mid-September Gallup poll that had Kerry 14 points behind Bush. The problem was, the poll assumed that on Election Day 40% of those casting a ballot will be Republicans and only 33% will be Democrats - a turnout that will happen only in Karl Rove's dreams. Democrats accounted for 39% of those voting in the last two presidential elections, while Republicans accounted for no more than 35%.

With computers, it's garbage in, garbage out. With polls, it's faulty data in, faulty findings out.

Yet polls are now firmly entrenched as the lingua franca of political analysis. Dissecting the latest numbers is so much easier than actually digging for the truth. Cable shows love turning campaigns into horse races. And it's so much easier if you can parade fatuous numbers as hard facts to prove Who's Hot and Who's Not.

Trouble is, these snapshots of the electorate quickly harden into portraits and have a profound impact on a campaign. The sense that a candidate is tanking - or on a roll - can make the difference between a potential donor making a contribution or keeping his checkbook in his pocket. It can also tip the scales for a would-be volunteer deciding whether to give up more free time to go door-to-door registering voters or work the phones to get out the vote.

This is how polls morph from meaningless farce into potential tragedy - self-fulfilling prophecies that end up making more likely whatever results that they predict while at the same time undermining the democratic process.

Add in the fact that pollsters are finding it increasingly difficult to get people to talk to them. Thanks to answering machines, caller ID and telemarketers, polling response rates have plunged to 30% and lower. It's pretty hard getting a good read on the public's opinion when people keep hanging up on you.

Plus, pollsters never call cell phones, of which there are now close to 170 million. And even though most cell phone users also have a land line, a growing number don't - especially young people, a hard-to-gauge demographic that could easily turn out to be the margin of difference in this year's race.

Our media mavens know this but choose to ignore it. Coming clean about polls would mean taking them off the front pages and sticking them where they belong - back among the horoscopes and comic strips.