[kevkrom]

This is the first time that I've heard that Rasmussen uses a four-day average. I've looked at the information on their site and it doesn't say this anywhere that I've seen. Can you provide a source for this information?

I know I've told you that before -- that's why I suggested using a four-period moving average -- but nevertheless, at the bottom of each daily article, they add the following boilerplate (relevant info highlighted):

Daily tracking results are from survey interviews conducted over four days ending last night. Each update includes approximately 750 Likely Democratic Primary Voters and 600 Likely Republican Primary Voters. Margin of sampling error for each is +/- 4 percentage points with a 95% level of confidence.

[Spiff]

I know I've told you that before -- that's why I suggested using a four-period moving average -- but nevertheless, at the bottom of each daily article, they add the following boilerplate (relevant info highlighted): Daily tracking results are from survey interviews conducted over four days ending last night. Each update includes approximately 750 Likely Democratic Primary Voters and 600 Likely Republican Primary Voters. Margin of sampling error for each is +/- 4 percentage points with a 95% level of confidence

Thank you for the information. I've read that blurb at the bottom of the Rasmussen commentary, but never interpreted as you have. I interpreted as each poll was taken over 4 days, not that they were reporting a 4-day running average.

Nevertheless, you're right. The full effect of the NRLC endorsement has not been seen yet as the posted results include responses from days ago, before the NRLC announcement was made.

[Quicksilver]

Having thought about it some, I don’t think that it’s an average. Taking what it says literally, for Republicans each daily result is the actual percentage of the most recent 2,400 samples. Tomorrow’s result will have the oldest 600 samples replaced with today’s 600 samples.