Since it became clear sometime during the past few weeks that the fate of health care reform rests in the House, a number of journalists and other political observers have begun tracking the “whip count”—the number of House Democrats who can be expected to vote in favor of the Senate bill, on the understanding that amendments to the bill would subsequently pass both houses. Efforts in this vein can be found here, here, here, and here, and the list goes on.

From one perspective, this makes a lot of sense: after the endless messaging and speechifying of the past year, we’re in the health care endgame, which is all about the votes. And the way to know where the votes are is to count them by tracking what legislators say about their stance on the bill. That’s as straightforward as it gets.

Except, maybe, it isn’t. The Washington Post’s Ezra Klein, whose blog has become one of the go-to sources of health care coverage for many readers, explained yesterday why he doesn’t trust these whip count features:

To a degree that’s really under-appreciated, legislators communicate with each other through the press. That’s particularly true for the specialized Washington press: The Roll Calls and Congressional Quarterlys and Hills of the world. We read those stories as news articles, but the politicians who are quoted in them are often directing their words at their leadership rather than the publication’s lay audience. And that means those comments generally have a couple different meanings. The most common double meaning is that a quote that seems like a prediction (“I am not inclined to vote for this bill”) is actually the opening of a negotiation (“I want something changed in this bill, and I want it badly enough that I’m willing to play up my opposition in the press.”)

As Klein notes, this may be happening with Bart Stupak, the pro-life Michigan Democrat who has caused all manner of headaches for his party’s leadership with his insistence on restrictive language regarding abortion. If even Stupak is making noises about compromise, it’s a sign that the votes may be more in flux than seemingly diehard public pronouncements suggest.

But if we can’t trust whip counts to make sense of likely outcomes, what can we do instead? One approach is to look at the incentives facing legislators. Several writers, including Time’s Karen Tumulty, have aptly summarized the situation: Democrats, as a group, have an interest in passing health care reform. But individual Democrats from swing districts have an interest in voting against reform; their ideal outcome, politically speaking, is to have reform pass without their vote. It’s easy to see how this could lead to stalemate; Tumulty analogizes it to the thought experiment known as the prisoner’s dilemma. But in similar episodes in the past, party leaders—yes, even those ever-in-disarray Democrats—have been able to extract the deciding vote.

Another alternative is to look for guides or signals. In the wake of Scott Brown’s election, health care obituaries were easy to find, and even Democratic leaders were indicating they might scale back the effort. But more recently, both the White House and leaders in Congress have been behaving in ways that suggest they believe comprehensive reform will get the votes it needs (even if they still have to pin down exactly who will be casting them). They may be bluffing, or they may be wrong—but they have some pretty strong incentives to be right, and also, presumably, access to good information.

What’s interesting about this, journalistically, is that it forces reporters and other observers to make decisions about which sort of information they’re going to prioritize. That’s a calculation that happens all the time, of course, but it’s often not apparent to readers—and, in some cases, may not even be apparent to journalists—so kudos to Klein for making his thinking explicit.

So which approach is better? The whip count angle is clearly more traditionally “journalistic”—documenting the public statements of elected officials is one of the things journalists do, whether or not they’re predictive—but Klein (echoed by Jon Bernstein) raises some real concerns about its utility here. On the other hand, while history, structure, and incentives are all invaluable in terms of understanding the general state of play—and while, in general, reporters should pay more attention to them—they’re not deterministic of the outcome in any particular case. In the end, while we all want to be the first to figure out what’s going to happen with health care, it makes sense to keep in mind something else Klein says in his post: we’ll know soon enough.

Greg Marx is a CJR staff writer. Follow him on Twitter @gregamarx.