Okay, okay. Maybe mid-September is a little early to be talking about such things, but I’m beginning to pick up a sense that the prevailing script in the news media – that the Democrats are poised to upset the balance of power in Washington – may have been over-hyped over the past few weeks.

I realize politics is an up-and-down thing and that momentum can shift more than once even at the end of an election season, but here’s why this sense of an impending Democratic takeover seems to me to be overreaching:

Parties are secondary. The polls that count are the ones with candidate names on them, not party preferences. Generic polls that test the electorates’ comfort level with the Democrats and Republicans in Congress may be useful in understanding the background noise for this year’s elections, but voters don’t face off between an elephant and a donkey in the polling booth – they choose between two individuals.

It’s one thing for a voter to tell a pollster they don’t like the way the Republicans have run Congress and would rather see the Democrats in control. It’s quite another thing for them to toss out their GOP incumbent for the particular Democrat whose name appears on the local ballot.

The best example of this I can think of is the Senate race in Pennsylvania, where GOP Sen. Rick Santorum is facing off against Bob Casey Jr., son of the state’s popular ex-governor. Santorum’s hard-right style of politics has not gone down well in moderate Pennsylvania, and Casey started out of the box with a double-digit lead. But that was before Santorum started fighting back – and before Casey’s performance on the stump revealed him to be a less-than-thrilling campaigner. If the election had been held last month, Santorum would have lost. But Election Day is in November, which brings us to the second factor…

August doesn’t count. As George Bush Sr. can personally attest, voters’ attitudes on Election Day are more likely to reflect their sense of things in November, rather than appreciation for things done in the past – even the recent past. Lots of red-blooded Americans don’t focus on politics even after Labor Day, let alone before it. So polling numbers from the dog days of summer are always suspect.

Likewise, because the voters are not focused that early, the political pros that run campaigns always save their heavy artillery – especially the powerful negative artillery – until just before the election. By August, we may know the names that will be on the fall ballot. We may have seen the candidates in early sparring. But we haven’t seen the bulk of the ads, and make no mistake, those last-minute ads can move things. After Labor Day, the game changes big time.

Everybody loves a drama – especially the news media. This may be the biggest reason why the Democratic surge has gotten so much ink in the past couple of months. Who wants to write about a ho-hum election? In this respect, news reporters – and the organizations that employ them – are just like everyone else. They’d rather have drama than no drama, so any scraps in the wind – like generic polls – that can be interpreted as harbingers of serious change are thoroughly reported. Any counter-news, like fundraising totals for incumbents, tend to wind up on the back pages rather than page one.

There’s been no end of commentary on the perceived liberal bias of the news media (and on counter-arguments that it’s not true). But my own sense is that the drama bias is much bigger than any political bias, and few journalists or news organizations are exempt.

Watch the money. What really got me on this line of thinking today was a news story in Roll Call by Alexander Bolton. The headline: “Mehlman to spend $60M, five times Dean’s $12M.” In short, the head of the Republican National Committee, Ken Mehlman, is poised to spend all $60 million of the RNC’s reserves on the mid-term election – five times more than the $12 million that’s available to DNC chairman Howard Dean.

Not all the money in this election is coming from the national parties, of course. Much larger sums have been raised by the candidates themselves and by their fundraising committees in the House and Senate. But a $48 million disparity by the national parties is nothing to sneeze at, and $60 million can buy a lot of negative ads.

That’s important, because the one safe bet in town, whichever way the momentum swings between now and the 7^th of November, is that this is going to be one down and dirty election. If the Republicans can once again outspend the Democrats, that money will do a lot to erode voters’ confidence in the newcomers seeking to updend the current regime.

What it may boil down to in the end is this: in competitive districts where the Democrats have offered up first-class candidates in tune with the electorate, skilled at campaigning, and competitively financed, they’re well poised to win in what should historically be a Democratic year. But in states and congressional districts where any one of those ingredients is not in place, 2006 is going to be just as rough a year as any for the toughest job in American politics: overcoming the inertia of incumbency.

The Democrats may still pull it off, but I’m not taking any bets just yet.