tag:blogger.com,1999:blog-6719252574677567989.post3152409555791353007..comments2024-03-26T05:22:08.256-04:00Comments on Frontloading HQ: The Electoral College Map (7/20/08) [Update]Josh Putnamhttp://www.blogger.com/profile/06301836432446874997noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-6719252574677567989.post-22656751216740917922008-07-23T11:14:00.000-04:002008-07-23T11:14:00.000-04:00Here's that RCP link from Rob.And here is the dire...<A HREF="" REL="nofollow">Here's that RCP link from Rob</A>.<BR/><BR/><A HREF="http://www.rasmussenreports.com/public_content/politics/election_20082/2008_presidential_election/ohio/election_2008_ohio_presidential_election" REL="nofollow">And here is the direct link to Rasmussen's report on the poll</A>. I'll have the new map up in a bit with more on Ohio. Hint, hint.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6719252574677567989.post-16228220647542242142008-07-23T11:09:00.000-04:002008-07-23T11:09:00.000-04:00I think this question needs more attention. I note...I think this question needs more attention. I note that Rasmussen's latest Ohio Poll has McCain ahead by 10% while other recent polls favor Obama.<BR/><BR/>http://www.realclearpolitics.com<BR/>/epolls/2008/president/oh/ohio<BR/>_mccain_vs_obama-400.htmlRoberthttps://www.blogger.com/profile/03379192575044761972noreply@blogger.comtag:blogger.com,1999:blog-6719252574677567989.post-86402702722914164752008-07-23T09:35:00.000-04:002008-07-23T09:35:00.000-04:00Scott,At the risk of sounding like a hack, I have ...Scott,<BR/>At the risk of sounding like a hack, I have made it a rule of thumb to follow everyone else on this type of issue. My thinking here is that it should be the methodology that is unique in our electoral college analysis, while the underlying data is as close to everyone else's as possible. <BR/><BR/>Nate at FiveThirtyEight explained his reasoning on July 9 in a post called "<A HREF="http://www.fivethirtyeight.com/2008/07/leaners.html" REL="nofollow">Leaners</A>". My reading of that is that Rasmussen was running with the "with leaners" numbers before the switch to the new presentation. That happened around the first of July. The question then is "Why the change at Rasmussen?" And why push the formerly headline numbers on the backburner? That's a question for Rasmussen. <BR/><BR/>But you know what? We can look at this here. And why not? FHQ is a forum to explore this too. All the data (the links to them at least) are right there in our electoral college breakdowns. So let's look at those numbers and see what exactly the changes would have been. I might be inclined to shrug this off, but Rasmussen has provided a substantial amount of the state-by-state polling data recently, and it would behoove us, or anyone else with any lingering doubts about this switch, to check it out. So let's do that.<BR/><BR/>I've got the electoral college post and another post that I want to roll out today. If I have time I'll get this one out as well. One thing I've been thinking about as I've been typing this is that we could put this in with our end of month examination of the polling changes during July. I think this deserves its own post though.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6719252574677567989.post-56417854287363471022008-07-22T20:49:00.000-04:002008-07-22T20:49:00.000-04:00Speaking of bringing up the same issue on a variet...Speaking of bringing up the same issue on a variety of blogs...<BR/><BR/>Josh: I'm curious as to your reasoning for using the Rasmussen state numbers "with leaners." You do it, RCP does it, pollster.com does it, and 538 does it.<BR/><BR/>But Rasmussen himself does not. He treats the "without leaners" number as the headline, and compares it, even averages it, with earlier Rasmussen state polls that did not break out separate numbers for with and without leaners. This implies that the earlier numbers must have been without leaners.<BR/><BR/>I don't really have a big problem with your using either set of numbers. It will introduce a slight shift toward McCain on average, because leaners tend to go disproportionately for McCain. But that may be more consistent with the results of other polls, so it's just a correction to more closely reflect what you're trying to get at.<BR/><BR/>At 538, on the other hand, I've been jumping up and down trying to get them to notice that, since their model includes a "trend," it's inappropriate to compare old Rasmussen state numbers with the current numbers with leaners. That treats the half point or so adjustment as if it's <I>temporal</I>, and implies an ongoing trend rather than a one shot change in methodology. You don't have a trend piece like that, so it's fine other way. But I would like to understand your reasoning better, since almost everyone seems to be doing what you're doing.Unknownhttps://www.blogger.com/profile/14690577323454357276noreply@blogger.comtag:blogger.com,1999:blog-6719252574677567989.post-44059111227016544982008-07-21T17:04:00.000-04:002008-07-21T17:04:00.000-04:00Yeah, me too. This is the last time that I will a...Yeah, me too. This is the last time that I will allow this message to remain up on one of the posts here. It is one thing to add to the discussion, but to essentially shout the same thing over and over again isn't adding anything to what we're trying to accomplish at FHQ.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6719252574677567989.post-33328161046959211962008-07-21T14:06:00.000-04:002008-07-21T14:06:00.000-04:00This comment sounds familiar. I will refrain from ...This comment sounds familiar. I will refrain from repeating my challenge to it from the past.Roberthttps://www.blogger.com/profile/03379192575044761972noreply@blogger.com