Sometimes it does, and sometimes it doesn't. Sometimes a good day is replaced with an even better day. We don't know that until we actually see the next day's poll.
But it is extremely rare in Rasmussen's poll for any one candidate to poll 7 points (my calculation) to 11 points (somoene else's cauclation) better than the other candidate in a single day. This happens in all polls approximately 5% of the time. Something like this apparently happened in Obama's favor on Tuesday night in Rasmussen's poll. If it was a real trend rather than statistical noise, we would have expected to see the same thing in other contemperaneous polls, but no such thing happened.
If Tuesday was a +7 day for Obama and Rasmussen has Obama up +5 then the other two days must total +8 don’t they? If so, then the +7 point day dropping off may not make too much of a difference to the overall rolling average (it may drop O’s lead to +3).
For example:
Tue: O +7
Wed: O +4?
Thur: O +4?
Avg: O +5
Assume that once tuesday drops off it is replaced by an Obama +1 then the rolling average would be O +3. Where am I wrong?
If Tuesday was a +7 day for Obama and Rasmussen has Obama up +5 then the other two days must total +8 don’t they? If so, then the +7 point day dropping off may not make too much of a difference to the overall rolling average (it may drop O’s lead to +3).
For example:
Tue: O +7
Wed: O +4?
Thur: O +4?
Avg: O +5
Assume that once tuesday drops off it is replaced by an Obama +1 then the rolling average would be O +3. Where am I wrong?