What happened here?

A couple of decades or so ago, my five-year-old nephew was going from the porch to the living room, and didn't see the screen door. So he bounced off it, hit the floor, and paused for a moment. You could just see his mind working.

"What happened here?"

Then he burst into tears.

I didn't do any crying while watching the New Hampshire primary returns the other night, but I knew the feeling. What happened up there?

Oh, I know the way the returns went, as Hillary Clinton and John McCain were winners. (Food for thought that Clinton didn't get much credit for being the first woman to win a primary -- as compared to the reaction to Obama's equally history-making win in Iowa.) But the analysis afterwards was interesting too.

All of the polls said before the election that Barack Obama was headed toward a convincing win, perhaps by double digits. The candidates' polls even said that. But it was Clinton by three points or so when the counting was done. The polls were right on McCain's margin. So what happened to Clinton/Obama?

The spinmasters gave their usual explanations about the race, how women over 40 turned out for Clinton, how her display of emotion in a debate had warmed her up to some voters, etc. etc. etc. But that didn't explain what went wrong with the polls.

I had three thoughts. One, the independents messed things up. In New Hampshire, they can pick a primary and vote in it. The opportunity for mischief certainly is there. Could voters think Obama was headed toward a big win, and decide to help out McCain with a vote? Could someone leaning toward the Republicans think Obama would be the tougher candidate to beat and vote for Clinton? Hard to judge that. These open primaries sure can be complicated.

Then there's the theory that people lied to the pollsters. That one went unsaid until 1:45 a.m., when NBC's Chuck Todd (he's good, by the way) pointed out that often when elections go in bizarre directions, an African-American candidate is involved. (See Tom Bradley's run for Governor in California a quarter-century ago.) In other words, people told pollsters they would vote for an African-American, but it was a different story when the curtain closed in the voting booth.

Obama seemed to overcome a lot the week before in Iowa, but that was at a very public caucus. This was a different sort of electoral test. You'd hope we've moved beyond that sort of thinking, but it's at least possible.

Then there's the belief that a good-sized percentage of voters didn't make up their minds until very, very late in the process (18 percent on the day of the election alone), and they broke to Clinton in a disproportionate way.

Was there truth in any of these theories? All of these? Tough to say. It's impossible to know where the truth is. But we have learned a lesson before jumping to conclusions: count the votes.

Stop the insanity

Let me just review the Presidential selection process briefly, just to make sure I have this right.

A bunch of candidates spend about a year campaigning and raising money. The campaigning is mostly done in Iowa and New Hampshire, and fund-raising is done anywhere there is a hand to shake and some shrimp to eat.

Then the business of selecting delegates starts. Well, sort of. In Iowa, voters aren't selecting delegates, they are caucusing in order to pick delegates. Not many people understand the system, but somehow winners and losers are determined -- mostly on a basis of expectations. In New Hampshire, you can vote for anyone you want regardless of party affiliation. So if you are a Democrat who is scared stiff that Rudy Giuliani is the only Republican who can stop a Democratic win (and that's not a ridiculous thought), then you'll sprint to the voting booth and become an instant Republican.

And when we are done with Iowa and New Hampshire, we have clear front-runners as well as candidates who can no longer raise money and head for the hills, leaving a few buttons (we should have more buttons in politics, don't you think?) and signs in their wake. In a couple of more weeks, a few more primaries will be held, and the winners of the two major nominations may be determined.

And it won't be February. Which means the voting population won't be asked to pay attention again until ... un ... Labor Day. Seven months later. In meantime, 90-plus percent of the electorate won't have much to say about the outcome.

Does this make any sense? Do people really wonder why voter turnout isn't too high?

I like the idea of retail campaigning that Iowa and New Hampshire bring to the table. It's charming to watch the candidates work the crowds one at a time on C-SPAN. The two states seem to take the responsibilities seriously, too. It would be a great way to start the campaign. Instead, it feels too often like the beginning of the end.

I'm not sure what might work here -- a "no delegate picked before March 1" rule or something might not be too practical -- but this system makes the electoral college look downright civilized.