8537
VIP Member
You're right, it's not a complete circle jerk. It's just an almost complete circle jerk
It's certainly suffering from significant selection bias:
If you poll members of a rightwing free market think tank, it shouldn't be surprising that you get answers as you'd expect from a rightwing free market think tank. "independent" voters who previously took an FEN poll are obviously likely FEN supporters.
I don't think the bias here needs to be explained.
So, they post an FEN poll and try to claim that posting such a poll on "liberal and union" websites removes the sample bias? it doesn't.
not completely, but it shows that the sample is tainted to some degree... and although I agree with much of the results, it's because it supports my bias, but may not necessarily be an accurate sample. Therefore, my caution.If you followed the LINK and actually READ it, there are all the internals right there for you to see.
indeed! And they show an online poll relying primarily on FEN members and FEN member invitees.
IOW, it's a statistical circle jerk.
It's certainly suffering from significant selection bias:
Category 1 was a specific request to each of the 10,000+ poll-takers who had taken the FEN October 2009 pre-election poll. That poll had accurately forecast the lack of engagement by younger voters, the high turnout that would be seen by voters over the age of 50, and a strong move to the right by Independent voters, resulting in a predicted “blow out” by Republican candidates.
If you poll members of a rightwing free market think tank, it shouldn't be surprising that you get answers as you'd expect from a rightwing free market think tank. "independent" voters who previously took an FEN poll are obviously likely FEN supporters.
Category 2 consisted of a request to FEN members to take the poll.
I don't think the bias here needs to be explained.
Category 3 consisted of poll-takers who responded to ads appearing on conservative, liberal and union-sponsored websites.
So, they post an FEN poll and try to claim that posting such a poll on "liberal and union" websites removes the sample bias? it doesn't.