tag:blogger.com,1999:blog-5254275556127802599.post6129674753963018758..comments2023-10-24T10:53:32.288-04:00Comments on Rights of Way: PollingC Nealhttp://www.blogger.com/profile/07865122912479524567noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-5254275556127802599.post-27303663744830729672011-11-11T12:33:41.796-05:002011-11-11T12:33:41.796-05:00With the actual election behind us, it's amazi...With the actual election behind us, it's amazing the degree to which Mike Tipping's poll was accurate. He seems to know his way around constructing a poll.Wm. A. Everitthttps://www.blogger.com/profile/16476427699753661051noreply@blogger.comtag:blogger.com,1999:blog-5254275556127802599.post-55800959123233025372011-11-03T11:37:47.697-04:002011-11-03T11:37:47.697-04:00Lots of problems with this. Here are three quick p...Lots of problems with this. Here are three quick points:<br /><br />1. The category is other/none and it's likely that many or most of the respondents were actually undecided.<br /><br />2. Your description of how MoE works is incorrect. Here's a good primer: http://abcnews.go.com/images/PollingUnit/MOEFranklin.pdf<br /><br />MoE is different for the whole poll (the number usually reported, at a 50% result) than it is for individual results. The farther a result is from 50%, the lower the MoE. They also skew slightly away from the extremes - results are only right in the center of the error distribution when they're close to 50%.<br /><br />There's a very legitimate case to be made for including error bars in poll results, but yours are incorrect.<br /><br />3. No margin of error was calculated or expressed for the IRV extrapolation. As the release makes clear, the stated MoE applies only to reported results for 1st, 2nd ad 3rd choice.<br /><br />For more discussion of methodology, see this follow-up post: http://www.downeast.com/the-tipping-point/2011/november/huge-response-down-east-mprc-pollMikehttp://mainepolitics.netnoreply@blogger.comtag:blogger.com,1999:blog-5254275556127802599.post-54035137346589870652011-11-02T20:46:12.531-04:002011-11-02T20:46:12.531-04:00Great analysis, thanks for crunching those numbers...Great analysis, thanks for crunching those numbers. The final results should be pretty interesting.Corey Templetonhttps://www.blogger.com/profile/16353022717289903546noreply@blogger.comtag:blogger.com,1999:blog-5254275556127802599.post-48139436382780336912011-11-02T15:54:18.009-04:002011-11-02T15:54:18.009-04:00Christian, I love your dorky mathy way!
Lisa F.Christian, I love your dorky mathy way!<br /><br />Lisa F.Lisa at Energy Circlehttps://www.blogger.com/profile/05709912230844372517noreply@blogger.comtag:blogger.com,1999:blog-5254275556127802599.post-91930373241032139042011-11-02T14:44:30.375-04:002011-11-02T14:44:30.375-04:00The precinct distribution is also totally out of w...The precinct distribution is also totally out of whack, over-sampling the islands and other precincts while under-sampling others.Anonymousnoreply@blogger.com