Hindrocket is no doubt feeling the effects of the long Minnesota winter, now nearing its midpoint. I took a somewhat closer look at the Post's current poll, and it looks like they're doing a Star-Tribune on the party registration balance. Correcting for this isn't enough to put Bush ahead, but it's certainly enough to close the gap by at least the stated margin of error.

Using Excel's Solver, I was able to tease out the reported party registration of the respondents: and it stands at Democrats 37%, Republicans 30.6%, Independents 32.4%. Because of rounding error, I had to run the solver for each response, and then average the results. It's obvious to even the most casual observer that these results make no sense. Giving equal weight to all three, which still probably slightly understates Republican strength, boosts Bush by about 1.5%, and cuts Kerry's support down by that much. It also puts the President's approval rating back up to 52%.

The raw data itself also has some anomalies. For instance, it has Kerry winning the South by 15 points, 55-40, his largest margin outside of the East, at 55-39. Now I can believe the East number, even in the Fall. But 15 points in the South? Can anyone think of a southern state Kerry is actually likely to win? What, did they poll Castro, project his opinions onto the entire island of Cuba, and then include Cuba? Even with the poll's stated error of +/- 3 points, the best you get is 52-43, which is still indefensible under the Laws of Common Sense.

The next step is to do the same for previous polls. Stay tuned.