tag:blogger.com,1999:blog-6926413038778731189.post8095255417431974262..comments2023-10-16T07:13:12.123-05:00Comments on A plain blog about politics: Read Stuff, You ShouldJonathan Bernsteinhttp://www.blogger.com/profile/15931039630306253241noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-6926413038778731189.post-47895547630273676232012-03-27T11:20:47.803-05:002012-03-27T11:20:47.803-05:00Kilgore may have been reasonable, but Silver is re...Kilgore may have been reasonable, but Silver is really just stirring the pot for no good reason here. <br /><br />My favorite model is Abramowitz's. That model is usually within 3 points, and it's very simple. It has both predictive and explanatory power. And yet, it's just casually treated as "one that isn't based on the fundamentals." Look at the models closely. You'll notice that, in the ones that include ANY kind of polling data (ie, all the ones Silver is only trashing by association and vaguely, as opposed to directly), the "fundamental" variables actually do a lot of heavy lifting. They do even more so once you step back and realize that polling results are, themselves, endogenous. They are CAUSED by the economy, war, etc. <br /><br />Next, Silver is going to tell the congressional modelers that they need to look at polling more, because it's more accurate than the fundamentals of the economy and incumbent presence in races. This is similar to how I can look outside and tell you that it's sunny now, and then when I actually get outside, gee, what do you know, I was right! As opposed to making a prediction based onn weather models. Yes, the models are less accurate....but I didn't need to know what the weather was today to predict the weather today!Matt Jarvisnoreply@blogger.com