The reservoirs of goodwill
by SevenStrings, Tue Aug 19, 2008 at 02:19:15 PM EDT
Some of you may recall a "reservoirs of goodwill" chart that I had occasionally presented during the primaries.
This metric plots the (%favorable - %unfavorable) and the (%very favorable - % very unfavorable) metrics for the candidates over time. I found that to be a useful barometer to gauge outcomes.
And so, here are the "reservoirs of goodwill" charts for Sen. McCain and Sen. Obama. The forced linear fits on the two sets of data points are just that ~ forced linear fits that are meant to guide the eye (the correlation is poor..blah blah...not enough statistics...blah blah..and I would not put too much faith on them)
The source of all this data is www.rassmussen.com
(a) Sen. Obama benefits from a low turnout.
Sen. Obama and Sen. McCain are about tied in the %favorable-%unfavorable metric (this is the first chart, which I call the "like him-dislike him" metric). However, Sen. Obama enjoys (or at least, he did enjoy) a substantial lead in the "love him-hate him" metric.
Thus, if voter apathy results in a low turnout, and only the highly motivated people (presumably, these are people who either love or hate one of the two candidates) show up to vote, then Sen. Obama would benefit.
(b) Sen. McCain benefits from a high turoutn.
This point follows directly from (a). I am highlighting it here because I have trouble believing it myself... I am sure you are sceptical of it too!!
Of course, turnout is a funny metric ~ a candidate wants high turnout amongst his supporters, and low turnout amongst the supporters of his opponent.. and it is not clear how voter apathy could affect that distribution.
(c) Sen McCan is definitely closing the gap on the "love him-hate him" metric
As of today, the "love him-hate him" metric for the two candidates are within 1 standard deviation ~ which means that the likelihood of them being equal is >65%
On June 18, Sen. Obama's "love him-hate him" metric was substantially greater than Sen. McCain's (the difference was 3x greater than the standard deviation).
(d) And yes, I understand the limitations of this analysis
This is a race for electoral votes, it is a state-by-state contest, one has to look for the polls in the battleground states, it is still early, Rasmussen uses automatic polling and their methods are suspect, Dewey beats Truman.. etc !!