In an
earlier diary entry, I analyzed the trend over several weeks of the difference in Kerry's and Bush's percentages on the Rasmussen Tracking Poll. That analysis revealed that, despite the noisiness of day-to-day fluctuations, in the long run Kerry was slowly and methodically pulling ahead. Four weeks have passed, and it is time to see whether or not that trend has held up.
Here is the plot of the differences (positive values mean Kerry ahead) over time to date:
[see earlier posting for details of analysis]
Clearly the trend has continued unabated. In fact, given the increased amount of data on which the trend is now based, the overall relationship is considerably stronger than before: the correlation coefficient of Kerry's lead with time is +.60 meaning that 36% of the fluctuations in the data can be acounted for by Kerry's progress over time.
Thus, while there continue to be ups and downs, it is apparent that, in general, the ups are getting progressively higher and the downs less low. It is good to keep this big picture in mind, especially since the local fluctuations are predominantly sampling noise and, thus, can be frustratingly unrelated to recent events (as is often bemoaned hereabouts).
Of course, it goes without saying that there is nothing causal implied by this analysis. Unforeseen dramatic events could certainly intervene to break the trend. But given the present environment and notwithstanding the considerable variation in the news on Iraq or the economy or the latest media preoccupation etc., Kerry's campaign strategy seems to be right on track.