While I do not get to see polls as often as most readers, I have observed that most polls have indicated something of a statistical dead heat (within the margin of error) for some time. Then out of the blue comes the LA Times poll [free registration required] showing Kerry with a 7-point lead in a head-to-head race with Bush. Anytime a poll is that far outside the mainstream of polling data, questions about methodology should be asked. In this case, Drudge, citing an article from Roll Call (which does not appear to be available to non-subscribers), reports that "Not counting independents, the Times results were calculated on a sample made up of 38 percent Democrats and 25 percent Republicans -- a huge and unheard-of margin." Given that skewed sample set, the only surprise is that Kerrys lead was only 7 points.