Back in
July and
August, I posted diary entries showing that, up to that point, Kerry was showing a slow but steady climb in the Rasmussen tracking poll relative to his opponent. A lot has happened in the meantime. Now, a week from the election, it is useful to return to that analysis to see where we've come... and, perhaps, where we are heading.
The plot below gives the difference in the candidate's percentages (Kerry minus Bush) since the beginning of February (You do remember February, right?) As before, the red curve gives seven-day running averages in the differences of the three-day average polled percentages that Rasmussen reports. The vertical green line on the right marks November 2.
As you can see, the previously reported linear trend provides a fairly good fit to the data for the first part of the plotted interval at the left side of the graph. Then, as we all well know, we started on a wild roller coaster ride!
The data for this later part of the race have been fitted here by a simple sinusoidal function (the blue curve), which does a remarkably good job in accounting for the trend. The region of the peak, of course, corresponds to the selection of Edwards and the Democratic Convention. The valley is the August of our despair: Swift Boats plus Republican Convention. The rebound from our lowest point is the period of the debates. These are admittedly quite distinct events whose timing was predetermined and whose effects on the voters was surely a result of the idiosyncratic particulars of those separate phenomena.
Nevertheless, a pretty pattern like a sine curve calls out for a simple unifying story. Is there such a story or is this making too much of coincidence? I don't know. But one can certainly think of at least a couple of plausible narratives.
One such story might be built on a theory of the sociodynamics underlying fashion and cultural trends. One could imagine concocting a diffusion network model with both mutual reinforcement and decay/fatigue processes that would lead to natural waxing and waning of activation patterns within the network. Oscillatory behavior can be the limiting state in such a self-organizing system.
A second and more compelling story would be based on John Kerry's mythic reputation as a brilliant politician. In this case, the pattern would be attributed to Kerry's understanding of the pace and rhythm of a political campaign: when to lie low, when to gather momentum, when to make the big kick that will carry you through to victory on election day.
Or one might even combine the two stories by saying that a master politician is one who is attuned to the natural resonance of the sociodynamic system. A pretty story, indeed.
Of course, all of this presumes that Rasmussen's numbers mean something in the first place (that is, that the trends in the numbers reflect something real). Still, I think looking at this big picture view captures something important in what we all intuitively feel about what we've been through. And the irresistible allure of that fragile blue line traveling along towards its fateful rendezvous with the finish line aptly represents why we're all so preoccupied simultaneously by our hopes and by our anxieties in this final week.