The conventional wisdom says we’re obsessed with polls, but I’ve always thought this was not quite so.

Yes, political junkies can scarcely wait to get their hands on the next hard numbers. That’s particularly so now, when we’re just starting to get the first real survey samplings for expected contested congressional races around the country. But what’s remarkable is our collective ability to ignore what the polls tell us when they don’t square with conventional wisdom.

Let me give you one example.

Back in 1998, I was just starting my career as a political journalist. I was writing for a small — and at that time even smaller — political magazine called The American Prospect, located up in Cambridge. I started in January of ’98, just in time to see the political world get turned upside down by the Monica Lewinsky scandal.

Later that year, I was — as far as I could tell — one of the relatively few folks who predicted that, far from getting blown out in the ’98 midterm, the Democrats would actually pick up seats.

I was, I’ll admit it, pretty impressed with myself. And various colleagues asked me afterward what I knew that they didn’t or hadn’t. But the simple truth was I just looked at the polls.

They were right. And they were right there for everyone to see.

The national generics weren’t hard in favor of the Democrats, but go back and look at the late numbers from ’98 and you’ll see they showed a modest but clear Democratic advantage. The problem was that most political insiders were so certain that voters would punish Bill Clinton and the Democrats for the president’s escapades that they had come up with all kinds of theories to explain why the polls weren’t really saying what they were saying.

The most common line was that whatever the polls said they weren’t measuring the sheer enraged intensity of the Republican base that year. That alone would push the balance back in the Republicans’ favor. Or perhaps Democrats would just be rendered too languid and inert by the shame of Clinton to ever make it to the ballot box. Or perhaps voters weren’t willing to share with pollsters the full measure of their revulsion with the president. Maybe that skewed the numbers.

Needless to say, pollsters have reasonably good techniques for gauging voter intensity and likelihood to show up at the polls — especially in the weeks and days just before an election. But pundits and analysts seem to forget that.

Now, it’s easier to remember these lessons if the polls are telling you what you want to hear rather than the reverse.

As many of you probably remember, just on the heels of the 2002 election, there were late-breaking polls that showed a final shift in the direction of the GOP. I remember discounting them in my own mind because they came so late and seemed to be belied by other data. But I do remember my own sinking feeling when I saw a Democratic pollster — who shall go nameless, unless you want to pull the transcript — on CNN’s “Crossfire.” Those polls, she said, simply could not be accurate.

Famous last words. And you know how that election turned out.

All of which brings us to the election we’re all focusing on less than four months in the future.

The polls right now look very bleak for congressional Republicans. Presumably they will slish and slosh this way and that before November, but we’re already starting to hear those same sorts of excuses and rationalizations for why those really bad congressional generic polls don’t matter.

Redistricting is too tight and efficient. The table is rigged.

Whatever the polls say, the Democrats have no plan, vision or message. So Republicans can breathe a sigh of relief.

National security rules in the post-Sept. 11 America, so unpopular as the Republicans may be the country won’t turn things over to the Democrats.

The Dems got crushed in the last two post-Sept. 11 elections. Why should 2006 be any different?

Each of those predictions contains an element of truth. Each may turn out to be the real story of the 2006 election. Polls aren’t perfect. You can’t see within the margin of error. And sometimes they miss something big.

But here’s my prediction: Watch the polls. Have your mind open enough to hear what they say. They’ll tell you the story.

Marshall is editor of talkingpointsmemo.com. His column appears in The Hill each week.
E-mail: This email address is being protected from spam bots, you need Javascript enabled to view it