What Do Americans Really Think of Bush's "Caller ID-to-the-Billionth-Power" Program? (And Why It Doesn't Matter)

If we needed one more example of why the results of political polls should be consigned to the same section of newspapers as the comics, Sudoku, and the daily horoscope, and not used as the basis for top-of-the-fold front page stories -- and, even worse, national policy -- we now have the contradictory results of two new polls purporting to show how the public feels about the NSA's massive collection of billions of our phone records.

Or, as I like to think of it, The Ultimate Pentagon Speed Dial (and you thought it was a drag when Paris Hilton lost her Sidekick).

On Friday, a Washington Post/ABC poll found that a very impressive 63 percent of Americans indicated that they were okay with the NSA collecting and keeping a permanent phone log on the calls made and received by tens of millions of Americans. I imagine the Bushies reacted as if they'd just heard that Stephen Colbert had been cancelled, and the Democrats began fretting about the president finally finding traction on something. For its part, Time magazine, inspired by the imposing result, put together a press release touting its new issue with the headline "Controversial Spying Program Could Give Political Boost to President Bush". (See how fast a "snapshot" becomes the Conventional Wisdom?)

Then, on Saturday, Newsweek (the Washington Post's corporate cousin) released a poll, according to which, the exact opposite was true: the majority of Americans thought that Bush's "Caller ID-To-The-Billionth-Power" Program was too big an invasion of privacy.

For good measure, on Sunday (it was apparently as busy a Mother's Day weekend for pollsters as it was for florists), USA Today/Gallup weighed in with another poll on the subject and, like Newsweek, found that a majority of Americans disapproved of the program.

The experts then stepped in to try and explain the differences: You see, the WaPo poll was done Thursday night, while the Newsweek poll was conducted on Thursday and Friday. Ah, that explains everything: People were obviously too busy watching Survivor and CSI to really pay attention when the WaPo pollsters called, but had fully boned up on the issue 24 hours later when Newsweek came a calling. Plus, the Washington Post was only able to convince 502 people to agree to take part in its poll, while Newsweek found 1,007 folks bored enough to oblige. USA Today split the difference and polled 809 adults. Then there was the usual microscopic dissection of the subtle differences between the wording of the questions asked: WaPo asked if the NSA data collection was "acceptable" or "unacceptable" while Newsweek offered respondents more detailed options on the program, asking whether "it is a necessary tool to combat terrorism" or "goes too far in invading people's privacy"?

I have a simpler explanation: most polls -- especially insta-polls like these, taken before people have had a chance to really think about the issue -- are garbage.

I've been writing for years now (here, here, here, here, and here) that the dirty little secret of the polling industry is that multiple factors are adding up to make polls about as reliable as Scott McLellan's pronouncements of Scooter Libby and Karl Rove's lack of involvement in Plamegate. These include plummeting response rates, and variables caused by the wording of questions, question order, the time the poll was taken, the weighing of sample groups for party affiliation, and that all-time favorite "the margin of error" (which doesn't actually mean what you think it does).

The real problem of course lies not in the polls themselves but in our leaders' obsession with following them and the media's obsession with reporting them.

Thus the vicious cycle of our poll-driven politics: Pollsters deliver their increasingly inaccurate polls; the media then report the results as if Moses had just brought them down from the mountaintop; and our poll-addicted politicians base their policies and pronouncements on the off-the-cuff telephonic responses of 502 people. Calls which the NSA has stored somewhere in its computer files. It's madness.

Of course, the larger point is that even if we knew with absolute certainty that -- with a margin of error of zero -- 63 percent or 73 percent or 93 percent or 103 percent of the American people were cool with the NSA keeping a permanent record of every single phone call made by every single American -- that still doesn't make it okay. Or legal.

Billmon nailed it: "The whole point of having civil liberties is that they are not supposed to be subject to a majority veto."

People can pick up their phones and vote Chris Daughtry off American Idol (and keep in mind, all you Katharine McPhee fans, the NSA has a record of who you are too). But they can't pick up the phone and vote off the Fourth Amendment. At least, not yet.

However the American people really feel about the NSA program -- and as more and more details come out, I have little doubt people will feel more and more uncomfortable with the idea of Big Brother having a record of who they called, when they called them, and how long the calls lasted -- this is most assuredly not a question of popular or unpopular. It's a question of legal or illegal, constitutional or unconstitutional.

And my distrust of polls cuts both ways. Democrats had better be careful not to be lulled into complacency by those comforting polls showing them with a double digit lead heading into the 2006 election.

They need to forget the polls -- whichever way they shake out -- and stay on the offensive. Harry Reid, Nancy Pelosi, and company need to remember: lead by the polls, die by the polls.

Follow Arianna Huffington on Twitter: www.twitter.com/ariannahuff