Comments on: Systematic biases from polling organizations http://redbluerichpoor.com/blog/2008/10/systematic-biases-from-polling-organizations/ Why Americans Vote the Way they Do Tue, 15 Dec 2009 10:59:42 +0000 http://wordpress.org/?v=abc hourly 1 By: Pollster biases revealed « Nachspiel at Polemarchus’ http://redbluerichpoor.com/blog/2008/10/systematic-biases-from-polling-organizations/comment-page-1/#comment-958 Pollster biases revealed « Nachspiel at Polemarchus’ Mon, 03 Nov 2008 10:26:26 +0000 http://redbluerichpoor.com/blog/?p=179#comment-958 [...] agree with Andrew Gelman who comments that this surprises him. I would expect their coverage of the polls to show some bias, [...] [...] agree with Andrew Gelman who comments that this surprises him. I would expect their coverage of the polls to show some bias, [...]

]]> By: evgen http://redbluerichpoor.com/blog/2008/10/systematic-biases-from-polling-organizations/comment-page-1/#comment-938 evgen Fri, 31 Oct 2008 20:40:24 +0000 http://redbluerichpoor.com/blog/?p=179#comment-938 With all due deference to Adelman and Schilling, they should leave polling analysis to people who know what they are doing. The fivethirtyeight analysis of pollsters lays out a good overview of the issues involved (including the devilishly difficult issue of "likely voter" models given the massive voter registration increase this year), but you can't help but laugh at a study which uses the RCP average as a baseline. The claim made in the paper: "We could not discover sufficient information about how Real Clear Politics calculates its average to justify a formal statistical analysis, however, our results are shown in the following table" should raise alarm bells for anyone who has followed the various polling sites this season. RCP is notorious for using poll selection criteria that raise suspicions of bias. With all due deference to Adelman and Schilling, they should leave polling analysis to people who know what they are doing. The fivethirtyeight analysis of pollsters lays out a good overview of the issues involved (including the devilishly difficult issue of “likely voter” models given the massive voter registration increase this year), but you can’t help but laugh at a study which uses the RCP average as a baseline. The claim made in the paper: “We
could not discover sufficient information about how Real Clear Politics calculates its average to justify a formal statistical analysis, however, our results are shown
in the following table” should raise alarm bells for anyone who has followed the various polling sites this season. RCP is notorious for using poll selection criteria that raise suspicions of bias.

]]> By: Andrew http://redbluerichpoor.com/blog/2008/10/systematic-biases-from-polling-organizations/comment-page-1/#comment-937 Andrew Fri, 31 Oct 2008 20:32:20 +0000 http://redbluerichpoor.com/blog/?p=179#comment-937 Phil, I don't know. But if you believe their analysis, then, as they point out, it's interesting that the biases correlate with the generally-perceived slants of the different news organizations. Phil,

I don’t know. But if you believe their analysis, then, as they point out, it’s interesting that the biases correlate with the generally-perceived slants of the different news organizations.

]]> By: Phil Groce http://redbluerichpoor.com/blog/2008/10/systematic-biases-from-polling-organizations/comment-page-1/#comment-936 Phil Groce Fri, 31 Oct 2008 20:24:20 +0000 http://redbluerichpoor.com/blog/?p=179#comment-936 <i>It still surprises me a bit–I thought all these pollsters were pros.</i> Should this imply, then, that you believe the biases are intentional? I don't see anything in the paper (I skimmed, admittedly) that would confirm that. Other places I read (e.g., fivethirtyeight), explain these in terms of differing underlying assumptions about the electorate. This seems reasonable to me, and doesn't suggest unprofessionalism, just a difference of opinion between pollsters. It still surprises me a bit–I thought all these pollsters were pros.

Should this imply, then, that you believe the biases are intentional? I don’t see anything in the paper (I skimmed, admittedly) that would confirm that.

Other places I read (e.g., fivethirtyeight), explain these in terms of differing underlying assumptions about the electorate. This seems reasonable to me, and doesn’t suggest unprofessionalism, just a difference of opinion between pollsters.

]]> By: Michael Weiksner http://redbluerichpoor.com/blog/2008/10/systematic-biases-from-polling-organizations/comment-page-1/#comment-934 Michael Weiksner Fri, 31 Oct 2008 19:26:39 +0000 http://redbluerichpoor.com/blog/?p=179#comment-934 This is very interesting. What makes this finding even more striking is that each pollster is aware of the other polls. So, they are aware that have a systematic bias as they correcting and then releasing their results. This is very interesting. What makes this finding even more striking is that each pollster is aware of the other polls. So, they are aware that have a systematic bias as they correcting and then releasing their results.

]]>