There’s been an outbreak of storytelling about the predictable decline of John McCain’s bounce and Sarah Palin’s favorable/unfavorable numbers.

For instance, Matthew Yglesias attributes Obama’s increasing numbers to the national focus on economic issues. Sam Wang calls it the “Palin Bounce.” Similarly, Ross Douthat blames the decline “primarily” on “negative press reports on her Alaska career… ongoing coverage of the still-simmering Troopergate scandal – and especially by her widely-watched, none-too-impressive interview with Charlie Gibson.”

From an epistemological perspective, we should be cautious about trusting any of these stories. The reason is that we have simpler explanations for both trends that have stronger empirical and theoretical support:

1. Presidential convention bounces tend to dissipate.
2. The poll numbers of national political figures become increasingly polarized when there is a two-way message flow about them (i.e. positive and negative information).

We can tell all sorts of (unprovable) stories about trends in polling numbers, which is what keeps pundits in business. At a minimum, it’s important to be precise about the claims that we make, keeping #1 and #2 above in mind. The relevant counterfactual for Yglesias is whether McCain’s bounce would have held up better had the national debate not shifted toward the economy. In the latter case, the question for Douthat is whether Palin’s numbers would have polarized as quickly if she had not been subject to as much negative coverage. Unfortunately, there’s no way to assess either of these scenarios.