Replies

Keep in mind that with the rolling average polls it is as important to look at the days falling off the average as it is the days being added to it. A change in the newest posted numbers might be the result of the impact of the day that fell off more than the day that was added. With the USC poll, we probably won't know the direction of momentum until the oldest three days fall off. Those days showed much more significant movement than the most recent four. My guess is, the USC poll will end up showing a pretty consistent +/- 1% lead for either candidate for the next couple weeks or until the next very significant news event.

One other lesson learned from the last two general elections...the realclearpolitics average is basically useless until it consists of all likely voter polls. Then it is a roughly useful indicator of who's ahead. But the actual percentage is only useful for looking at momentum over a period of days. Which frankly, is about all polls are really worth anyway, since the variables used to create them are inconsistent from one polling outfit to another. And Rueters has already proven this election that you can't even assume the polls of an individual polling firm are useful for trend analysis because they feel free to shift their polling strategy based on their desired outcome.