“Is participation in the poll by invitation only, or do the participants self select?”
They are asked if they would participate, and then once opted in they periodically issue questionnaire styloe polls.
Several demographic questions are asked each time: age, location, gender, party registration, how you last voted, how you would vote today.
As I said, they worked for years to refine the method. I see no reason why it couldn’t be just as accurate as phone calls.
Many people no longer use land lines. And many people don’t use the internet.
To correct for such things, they need to compare different methods, and learn how to make adjustments.
They all make adjustments, including Rassmussen.
When they all give the same result, it validates the trend.
Yeah, I agree it’s several cuts above the typical internet poll. I still don’t see it as being as credible as the traditional polls. I do agree it’s more evidence of a trend, though.