Keeping tracks

Chris Bowers over at MyDD points out that a gain in tracking polls does not mean the candidate had his best day last. He uses the Rasmussen poll as an example:

Even though Bush gained today, it was not because he had a strong single day sample yesterday. In fact, the only reason Bush gained today was because Kerry's strongest single day sample in five weeks, Monday's 48.0-44.1, dropped out of the three-day average. Kerry actually has an early lead in Sunday's three-day average, 15.6-15.2. Kerry is slightly behind going into tomorrow's three-day average, 32.2-30.4.

Of course this is highly logical, and easily forgotten in partisan worrying about such a tight race. One bad sample skews the result for three days. All we know at this point is that the President did better today that four days ago, and there is no way to know if that will e sustained until tomorrow.

Now this morning MSNBC reported on John Zogby's analysis in his latest tracking poll. Here's what they quoted:

The good news for the President is that he has improved his performance among the small group of undecideds.

Here's what they left off:

Nearly a quarter now say that he deserves to be re-elected—up from 18% in our last poll.

Somehow I doubt undecideds breaking 3:1 for Kerry would be "good news" for the President. But that's just me.

Keep an eye on tomorrow to see if these trends continue.