Republicans have been pushing hard this week to convince people that Mitt Romney is wrapping up the presidential election. Since he’s not actually, well, leading, Romney partisans have relied on the idea that Romney has momentum: Even if he isn’t actually ahead yet, he is certain to take a commanding lead any minute now.

But that “momentum” appears to have been entirely an invention of Republican spinners. It’s certainly true that Romney made impressive gains on Barack Obama in roughly the first week of October, probably in most part as a consequence of the first debate. But after that, the contest has been almost completely flat. For example, the Pollster trend line shows the race a dead heat on October 8 — and that since then, any movement has been only by small fractions of a percentage point. Nate Silver’s “nowcast” bottomed out for the president on October 12, and since then he’s recovered quite a bit. There’s simply nothing in the last twelve days to indicate movement towards Romney

Putting aside the present spin: why, indeed, should we ever talk about momentum? In fact, in thinking about elections, the idea of momentum is useless.

It’s worth thinking about three groups of voters who might change the horse race polls.

The first group is made up of truly independent, and undecided, voters. Most of them don’t pay very much attention at all to politics. Eventually, they will vote one way or another (at least if they vote), but they may be the most likely to swing between “undecided” and that candidate depending on what’s currently in the news. This create the impression of momentum, but the effect will be short lived, rather than continuous.

The second group is made up of voters who have actually decided but move in and out of “likely voter” screens, or even in and out of answering surveys at all, depending on their enthusiasm for the election and their candidate at any given moment. This can create a temporary “surge” in likely-voter support for one candidate that isn’t actually long term momentum in his or her favor.

The third group is made up of so-called undecided voters whose ultimate decision is basically predictable. Thanks to the campaign, they end up “learning” to vote for the candidate they might have expected to support all along. For example, there are those voters who insist that they are independent, yet always just happen to wind up supporting the same party. For them, the high-intensity portions of the campaign in which they are exposed to “their” party’s message will often be sufficient to permanently move them from undecided to decided. That can create a bump for a candidate — and the appearance of momentum — but once it’s done, it’s done. There’s no reason to expect it to build on itself.

In all these three cases, shifts happen when the information environment strongly favors one candidate or the other — such as when a party’s (successful) convention is running, or if a candidate is perceived to have decisively won a debate. The key thing, then, is that “momentum” is only likely to last as long as information favoring one candidate continues to dominate the news. And that almost never happens, because the press usually wants to move on to a new story.

For the first two groups, it’s likely that the effect will wear off rapidly once press coverage shifts, as it almost always does. By contrast, for the third group, some of the gains from a “momentum” period are likely to be permanent conversions. But they would have ended up with that candidate, anyway.

All of which means that “momentum” is the wrong word to use, since it implies a continuing trend — one that causes further gains. We’re better off thinking of “bumps” and “bounces” — reactions to events, which move a set of voters for either a very short time or a longer term. In Romney’s case, he had a “surge” in early October — but that tells us absolutely nothing about which direction the polls are moving in now and will move in until election day.

The real question isn’t so much whether Romney — or Obama — has “momentum,” but whether there is any such animal when it comes to presidential elections. For the most part, it’s just a term campaigns use to excite their partisans and to fool gullible reporters into writing stories that create the illusion of momentum that never existed in the first place.