(This is the underlying set of post which are set to kick off a discussion about Wittgenstein, connoisseur judgment and statistics next week). Greetings Tracy. Replying to this short summary of Holbrook: "presidential approval ratings don't tell us much of anything about how presidential elections will turn out." I'm not convinced. That analysis doesn't tell us anything. The appeal to the regression only creates a false charm. I wonder what would happen if you modeled it with a binary outcome (win loss) and only entered cases of 40 or below, or 45. (It's a small n either way). The theory isn't that approval is one-to-one correlated with votes. The theory is that there is some basement-level of job approval that makes winning the election less probable. The theory is that low-approval ratings produce a different kind of presidential election dynamic. The idea is that undecided voters have more of an incentive to go with a challenger when the job approval of the incumbent is too low. That's what happened to Carter. Besides, although Reagan and Clinton dipped low, they had approval ratings that (overall) trended upwards during the life of their terms. Reagan's was particularly rocky (bouncing back twice), ending with a slight overall upward trend. All other presidents in modern times experience overall erosion, mostly gradual but some episodic. So, let's assume there is an erosion-trend president having numbers at 45 in July -- are you going to tell me that has no influence on the probability of winning the race? I don't think so. I'm not saying that Obama will lose. I'm not even saying he won't bounce back in the polls. And I'm surely not saying that what has happened in the past 6 other times has to happen again. But what I am saying is that it can't be good news. Truth is, I always thought Obama's charm and charisma would work a lot like Reagan and Clinton. I had always thought he would bounce back. But lately, I'm beginning to believe that it won't happen. And I fear that in July he'll have approval ratings southward of 44 and will be asking to become the first president since (we assume) before Roosevelt to have the incumbent party win with such job numbers. (Could anyone speculate who might have done this pre-Roosevelt, if we had the data?). My colleagues here think otherwise. They're all concerned with head-to-head numbers. But I'm concerned with late-breaking voters. Much like Carter, I fear that that the undecided vote 4 days before the election will break in favor of a challenger who represents an alternative to the perceived social misery. I hope I am wrong. Anyway, it's still witchcraft whether you use stats or not. That's the real thing that may be lost. The fact that someone uses a univariate 14-case linear regression doesn't make him or her less of a metaphysician on the subject. I prefer connoisseur judgment. You want to come to understand a cultural context (happenstance). In any event, thanks for pointing out the blog. I'm going to pay more attention to it now. Regards and thanks. Dr. Sean Wilson, Esq. [spoiler]Assistant Professor Wright State University Personal Website: http://seanwilson.org ; SSRN papers: http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=596860 ; Wittgenstein Discussion: http://seanwilson.org/wittgenstein.discussion.html [/spoiler] ----- Original Message ----- From: Tracy Lightcap <tlightcap@xxxxxxxxxxxx> To: Sean Wilson <whoooo26505@xxxxxxxxx> Cc: LAWCOURT-L@xxxxxxxxxxxxxxxxxxx Sent: Wednesday, January 25, 2012 5:25 PM Subject: Re: polls and elections Once again a visit to our friends at the Monkey Cage helps. See: themonkeycage.org/blog/2011/08/10/potpourri-elections-edition/ Then read the links to the short pieces by Tom Holbrook (very helpful graphs) and Drew Matthews. Holbrook more directly addresses this question. Short Tracy On Jan 25, 2012, at 4:59 PM, Sean Wilson wrote: > ... has there ever been a sitting president who had a Gallup Poll approval > rating of 45 or lower in the month of the first party convention, but before > the convention started, where the party in power won the White House? All I > can think of are general trends: (a) Truman's polls went very low, Ike came > in; (b) Johnson's numbers went very low, Nixon came in; (c) Nixon plummeted, > Carter came in; (d) Carter plummeted, Reagan came in; (e) Bush I fell, > Clinton came in; (f) Bush II fell, Obama came in. But I believe the numbers > in these cases go down in the 30s. > > Is there any precedent for a president having a 45 rating (or lower) in July, > before the convention, where the president (or that party) wins the White > House? > > Regards and thanks. > > Dr. Sean Wilson, Esq. > Assistant Professor > Wright State University > Personal Website: http://seanwilson.org > SSRN papers: http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=596860 > Wittgenstein Discussion: http://seanwilson.org/wittgenstein.discussion.html Tracy Lightcap Professor and Chair Department of Political Science LaGrange College 601 Broad St. LaGrange GA 30240-2999 (O) tlightcap@xxxxxxxxxxxx (H) altlamp@xxxxxxxxxxxxxx 706.880.8226 www.lagrange.edu/academics/political-science/faculty/tlightcap.aspx us.macmillan.com/thepoliticsoftorture _______________________________________________ Wittrs mailing list Wittrs@xxxxxxxxxxxxxxxxxxx http://undergroundwiki.org/mailman/listinfo/wittrs_undergroundwiki.org