Pitfalls Of Paying Too Much Attention To Polling

Doug Kahn is a new DWT writer and I'm trying to persuade him to write for us regularly. He ran against far right impeachment fanatic James Rogan down the road from where I live and he's been very active in Democratic politics since then, a major DCCC contributor, who has come to regret helping the DCCC elect Blue Dogs. He wants to help Blue America elect people who will fight for progressive legislation, and not fold up when the heat comes on. These days he's turning his attentions in the direction of progressive grassroots candidates like Doug Pike (PA) and Doug Tudor (FL)-- not because they're Dougs, but because they're the real deal. Originally he comes from the tiny Pennsylvania hamlet of Rose Valley and some consider him to be excessively cranky-- perfect for DWT. His first post for us:


I've been reading that the USAToday/Gallup poll has some people very downhearted about the chances of Congress passing reform of the health care system, and regulating insurance companies. Chuck Todd and the other Einsteins of journalism are jawing about what they believe to be an absolutely shattering revelation: that "34% of Independents are now more sympathetic to the views of protesters at town hall meetings." Overall, 34% of American adults have new, improved sympathy. This is the kind of news story known as "Dog Bites Man." That is, so what. The startling thing would have been if these numbers were any lower.

If you pay any attention to the poll at all [and I wouldn't recommend it, because of how it was conducted] you should be encouraged by some more important numbers. Despite the fact that 80% of Republicans are following the town halls very or somewhat closely, only half of Republicans have gained sympathy for the protesters' views. This drops off sharply among Independents and even more sharply with Democrats. Half of American adults don't know or don't care, and the protesters have pushed a fifth of adults away from the opposition to health care reform. This last item is worth examining. Negative views are easier to measure, and the numbers are more significant. Yes (and maybe) can mean a lot of things, but no means no. Positive ratings of people and policies are extremely volatile, but large negatives are difficult to budge.

It makes sense to think of the Republicans as the opposition to health care reform, and their negatives are already remarkable. In the latest Daily Kos weekly poll (released Friday morning) we have this about Independents' attitudes toward the Republican Party: 5% positive, 85% negative. Overall, 17% of Americans have a positive view of the Republican Party, 74% negative. The USA Today/Gallup poll has 20% of Americans saying they're going to dislike the Republican positions on health care reform even more. This is newsworthy. You have to ask yourself if it's really possible for them to go any lower.

For Republicans in Congress, the numbers are worse: 10% positive, 76% negative. However, ratings of Congress should be taken with a grain of salt. Ratings of 'Republicans in Congress' are skewed by the fact that people basically like winners and dislike losers; for instance, the losers of elections. Underdogs generally benefit from a certain measure of good will, but no one thinks of members of Congress as underdogs. The Democrats in Congress are disliked by Republicans (obviously) and by Independents (to a lesser degree), and by-- well, by people like me. Because I think many if not most of them are spineless and unprincipled, and not progressive enough. Measuring 'Congressional Favorability' also has a basic structural difficulty; it's important not to inadvertently slant poll questions, which can happen if you use certain words or phrases. Including the word 'Congress' in a question, while unavoidable, has a devaluing effect. It's like adding the word "leeches."

You don't believe me? Okay, what if I ask if you like Chocolate Ice Cream? How about Congressional Chocolate Ice Cream? Try it yourself. Congressional Porsche. Congressional Buffet. Congressional Honesty. And so on and so forth.

Now just to be completely open and above-board about the 34% figure, I have to inform you that Gallup is very confident, 95% confident anyway, that the true figure is somewhere between 38% and 30%. Also, the polling was done on one day only, last Tuesday, and this is usually a no-no in polling. There are big differences in who will talk on the phone on a Tuesday as opposed to a Saturday, and no one really knows how that plays out. Polls with a lot of questions, like this poll, further damage the random sample. More people will refuse to answer, these refusers are similar to each other, and will be underrepresented in the polling sample. You almost certainly miss many more young people and low-income people, people working two jobs, parents with small children, so you can safely assume the 34% figure is inflated.

The question was asked as part of the Gallup Daily Tracking Poll, following the many standard questions, which are asked each day of a thousand people, in the same order (to avoid methodological bias). Each question is obviously less important than the last. Look on the Gallup site. It goes from the standard Obama Job Approval question, to 'Is your place of work hiring or letting people go,' to 'How much did you spend on consumer items yesterday, in dollars'. Then 'Is the country's economic outlook getting better or worse'; 'Are economic conditions Excellent, Good, or Poor'; 'Is your standard of living getting better or worse'; 'Are you satisfied or not satisfied with the way things are going in the USA'. Next is a rating based on combined responses to ladder ratings from two components of the Cantril Self-Anchoring Striving Scale, which-- well, the point is the poll goes on and on, also 'measuring' Mood, and then Health.

After 10 or 15 minutes of dragging answers out of people, this question: 'From what you know or have read, have these town hall meeting protests against the proposed bills made you more sympathetic to the protesters' views, do the protests not make any difference to you either way, or have the protests made you less sympathetic to the protesters' views?' 48 words. This borders on statistical malpractice. You might as well ask people to diagram the sentence.

Compare Pirates of the Caribbean:

Sailor #1: "You've seen a ship with black sails that's crewed by the damned, and captained by a man so evil that Hell itself spat him back out?"

Sailor #2: "No. But I have seen a ship with black sails."

Sailor #1: "Oh, and no ship that's not crewed by the damned and captained by a man so evil that Hell itself spat him back out could possibly have black sails, therefore couldn't possibly be any other ship than the Black Pearl. Is that what you're telling me?"

Sailor #2: [Nodding yes.] "No."

So, you've been called by the most venerable polling firm in the United States, and you know you've been chosen to represent 300,000 Americans. At first, you're asked the kind of questions that indicate you've been judged capable of intelligently analyzing political, economic, and social issues. But it starts going downhill. Soon you're being forced into a choice of describing yourself as "energized," or in the alternative, "ailing."

Finally, they ask you about the health care debate, and 69% of us are paying attention to this debate, with 32% very close attention. But they don't want your opinion. No, they want to know if a circus act on TV has made you more or less sympathetic (whatever that means) to some very loud people who are definitely on, shall we say, an alternate intellectual vector. WTF? So, if for 23 hours out of 24 you really dislike these people, but now it's only 20 hours out of 24, the answer is yes. Or, maybe you like these people, they're kind of like you, and now you really like them. That's also a yes. It's like this: my friends have heard me describe myself as having had 'mixed feelings' about a certain former campaign manager, because half the time, I really, really, you know, disliked him. The other half of the time, in contrast, I wanted to put my hands around his throat.

Sympathetic to the protesters' views? You mean positive about the views, or do I sympathize with the person? Responses to this question simply can't be analyzed scientifically. For instance, I sympathized with a crying woman at one town hall meeting, who was obviously distraught, and I truly wish I would hear that someone had managed to help her calm down and feel better. But her views were absolutely irrational. "Give me back my country," I think she said.

See if the following scenario doesn't describe this type of polling better than the word "scientific." Imagine you're in a room having a rational debate (okay, you're arguing) with 5 other people about the health care bill.

One of them is opposed just because he's a Republican and you're a Democrat; this guy is likely to say he's more sympathetic to the protesters' views, now that he's seen them shouting down Democrats at the forums. In fact, he's thrilled to death about it. So that portion of the 34% makes sense, it's logical. The other Republican isn't more sympathetic; maybe he's one of the 58% of Republicans who think the people who are shouting down supporters of the bill are abusing democracy. Of course, 51% of Republicans say they're now more sympathetic to the views of these protesters. And about 10% of Republicans haven't been paying attention at all. That's 119%, and still counting. (If you ever run for office and walk precincts, talk to voters at their front doors, you'll learn very quickly that diverse and seemingly contradictory opinions co-exist in very reasonable people. Anyway, that's what a lot of them said about me. And I was thinking the same thing about them. I can honestly say that I never met two people with exactly the same combination of opinions.)

One's a Democrat who is less sympathetic. This person hasn't really been paying much attention until the past week, but now she has people from various groups emailing and calling, trying to enlist her in cajoling her Representatives to the obvious conclusion that something needs to be done. ["Shouldn't they have been paying attention before 2009?"] She's been working 10 hours for 8 hours pay, schlepping groceries home or kids to school or her professional colleagues to conferences or writing briefs nights and weekends because she'd really like to make partner at the law firm, but now she's thinking, "where have I seen this before," because it's become clear from video that the protesters are mostly angry old white guys who enjoy yelling at other people. This has really been building up in these particular seniors since election day last November, when their guy got shellacked. Your average, totally burned-up, "had just about enough of this shit" mouthing-off male is not unlike the loser in that election; our loyal Democrat can picture him in his folding lawn chair scaring the neighborhood kids. Any more aggravations and you just know he's going to throw his scotch and soda across the lawn. The protesters are frosted because the winner is just not their type of guy, if you know what I mean, and also a foreigner. She's mad because Doofus McCain picked a woman as his running mate, but this VP candidate didn't exactly elevate public admiration for professional women; she was someone whose only skill was knowing how to win popularity contests. Unless you consider using sarcasm as a method of rational discourse to be a skill. (If your kids are older than 6 or 7, you've stopped thinking that. However, you get a free pass when you're blogging.) So this portion of the 'less sympathetic' category isn't hard to understand, either.

You're the other Democrat, and you're not more sympathetic, but you're also not less sympathetic. After what's been going on over the last 40 years, you couldn't be any less sympathetic. Not possible.

Your other two debaters are Independents. Two-thirds of the voting population consists of core Democrats and core Republicans. 60% to 70% of voters won't cross party lines, so Independents are the key to elections. Understand what that means. They can (and did) vote in overwhelming numbers for Richard Nixon and Ronald Reagan. But you never know; they might decide to pitch in for Al (Brainiac) Gore, or Barack Obama. Richard Nixon and Barack Obama! What could be going on in their minds, exactly? I mean, we all know it's sensible to be flexible in our thinking, but these people are mental bungee-jumpers.

Your Independent category includes members of the other umpteen parties, who get lumped in with people who are 'aginners, that is, those who lean one way or another but are not about to admit it in writing. I myself live in Florida, or as some of us have taken to calling it since the Al Gore debacle, Flori-duuhhh. One of the major political groups down here is the Secesh (as in secession) faction, or what we used to call Crackers, and their thinking is truly Independent of their neighbors (who in contrast are mostly residents of the planet Earth). If you grew up with them in the 1960s you've been trying to erase the memory of this classic bumper sticker: Hell no, we ain't fergettin'! These geniuses have matured in recent decades to more subtle affectations, like Silhouette mud flaps.

Some of the other parties you've already heard of. The Green Party. Ralph Nader, a great man who almost single-handedly broke through the massive corporate lobbying against requiring new cars to have a seat belt for drivers. However, there's such a thing as too sharp a point on a pencil. Greens wouldn't be sympathetic.

The American Independence Party (Nazis), sympathetic.

Peace and Freedom Party members, who really are, not kidding, just about the nicest people you'd ever want to meet. Very sympathetic to the protesters, but they'd also empathize with the House members hosting the forums, and do you have any doubt they are all for Single Payer?

Libertarians. I don't know what to say. One of Ron Paul's yard signs from last fall was "Choose Freedom." It seems he's in favor of freedom.

So, six people in the room. Two are more sympathetic now, and at least one of these will vote. In 14 months. Are we still worried?

Back to the poll. Please think about the 34% of Independents who are now more sympathetic to the protesters. Thinking about the protesters you've seen yelling, do you think many of them are registered Democrats or Republicans? I don't. I'd say upwards of 90% of them are Independents. 35% of the electorate is Independents. 34% of 35% is 12% of the electorate.

Yes, around 12% of the electorate is elderly, white wingnuts who don't trust either party, and who no one else in the country pays much attention to. (For instance, this describes my parents. Anyone else having the same experience?) They think having Barack Obama as President and Nancy Pelosi as Speaker, and the fact that most people finally realize that LGBT people have been shamefully discriminated against: it means the country they grew up in is gone. Yes! Thank God!

"Give me back my country." No. Sometimes you just have to say no.

I know this isn't particularly relevant to the health care debate, but I have to say I love the Gallup 'Mood' question, which proves the startling truth of both TGIF and Blue Monday. Every week without fail the chart is a sawtooth from 58% happy to 41% happy, Friday to Monday and then back again. Anomalies: happiness reached a peak of 67% on November 27th, 2008, Thanksgiving Day. Christmas Day was second at 65%, no doubt slightly reduced by the number of parents who were assembling bicycles manufactured in China, plus those who don't celebrate Christmas and feel left out. Easter was next at 64%, then New Year's Day, at 61%. Only about 10-15% report feeling really crappy about things on an average day, which surprises me. If you checked on me at random intervals, I think I'd report significant crappiness more than 15% of the time. Maybe if I stopped reading polls.




UPDATE: A Perfectly Legitimate Use Of Polling-- Comedy

OpenLeft demonstrates the value of polling the unknown, inspired by a poll of North Carolina Republicans who claim that Obama isn't really president because he was born in Hawaii which either isn't or might not be part of the U.S.