September 07, 2004

Poll Histories: Kerry and Dukakis

Instapundit has been collecting Dukakis-Kerry parallels, e.g. here and here. This is, of course, a fun subject for conservatives at the moment. My electoral college map and prediction are looking better all the time, though now I'm starting to wonder whether maybe Bush will take California (see also

(for a better display, click

Here's my take on Kerry versus Dukakis. The similarity is that both were extreme liberals facing somewhat unpopular moderate conservatives in good economic times, but both were perceived as moderates and the press, heavily on their side, did all it could to preserve the moderate image. Both therefore had a strategy of not talking about issues, but that strategy ran into trouble when their opponents adopted the simple expedient of pointing out their liberalism using ads with specific, undisputable examples. At that point, their advantage in the polls started evaporating, as the diagram in this post shows. I've noted in a previous post the Rasmussen Reports polling data that shows that middle-of= the-road (i.e., wishy-washy) voters wrongly think Kerry is not a liberal. Kerry has a lot of support to lose still, when they learn more about him. Note what must be disturbing for him: Dukakis was in much better shape than Kerry until August, when they both started declining, but by September 1, Dukakis and Kerry were about even. Dukakis then fell even further-- and he didn't have Swiftvets to worry about.

Soxblog talks about "DUKAKISIZATION": how a candidate becomes a laughingstock. Kerry is an easy target for that, unlike Dukakis, who wasn't particularly goofy in appearance and manner until his campaign people told him to do things like drive tanks.

There are two big differences, I think. First, Kerry has a foreign policy records that can only be described as pro-communist, while Dukakis, a governor, had no foreign policy record.

Second, Kerry is highly vulnerable on character issues. His military record is fraudulent. He lied about American atrocities when he returned to America. He won't release his wife's income tax returns, an unprecedented secrecy in the recent history of presidential campaigns, which tells me there is something rotten hidden there (though it might be as simple as, say, not having any charitable deductions). He divorced his first wife under conditions as yet unexplored. He has lived extravagantly using the money of his second wife's dead husband. I don't recall anything in Dukakis's personal life that was as distasteful as any one of these Kerry defects.

What can Kerry do? If he wants to win, I think his best strategy is to keep on obfuscating and denying, and hope that somehow Bush self-destructs. That strategy is not likely to win, but that doesn't mean it isn't his best strategy. He is just not playing a strong hand. Or, if he wants to serve the public, he could imitate Goldwater and say all the extreme things he really believes, in the hope of moving public opinion in the long run even though he would lose the 2004 election.

It's time to bring back an old joke, I think, though it doesn't work as well as it did for Dukakis:

QUESTION: What does "Kerry" mean in the original French?

ANSWER: "McGovern".

By the way, on Dukakis vs. Kerry polling trends see also the Free Republic of July 17. One interesting comment there was:

I think another KEY point is the MEDIA BIAS factor, where polling data during the 'non-attention' time of the voters

Can someone run an analysis of the July/August #s versus the final numbers?

I think if you go all the way back to 1976, in every single race, the
Republicans improved on their July/August polling numbers:
- 1976 (large Carter lead -> small Carter win)
- 1980 (Reagan behind/even race -> big Reagan win)
- 1984 (small Reagan lead -> Reagan landslide)
- 1988 (Bush behind/Dukakis ahead -> Bush wins)
- 1992 (Clinton lead -> Clinton win smaller than Aug lead)
- 1996 (Clinton large lead -> Clinton wins by 7pts)
- 2000 (no change? Bush/Gore even in summer, then close election)

This is something worth exploring, perhaps even in a scholarly paper. Gallup has the data at the site I link to below.

Here is the explanation for my graph above. It uses Gallup poll data, recorded a bit sloppily because each date is actually from about 3 days of polling taken from here and here from I converted calendar dates to digital dates, and my Dukakis data is eyeballed from a strangely drawn Gallup graph. The poll is of "likely voters", in a two way race plus Neither, Other, and No Opinion, where I eliminated the Neither, Other, and No Opinion and show a candidates proportion of what remains. I graphed the data in STATA using the command

graph kerry dukakis date, xscale(3, 11.5) yscale(40, 60)xlab(3,4,5,6,7,8,9,10,11) ylab( 40, 45, 50, 55, 60) yline (50) connect(l[.] l[l]) symbol([kerry] [dukakis]) psize(200)

Then I saved it as a *.wmf file and converted to *.jpg. I couldn't figure out how to keep the Stata colors when I saved it, though.

Posted by erasmuse at 11:01 PM | Comments (0) | TrackBack

August 28, 2004

Electoral College Map; Campaign Finance Reform

UPDATE, SEPTEMBER 5. See also the Gallup interactive state map, which has links to state polls.

The PBS Electoral College Pick-Your-Own-States Map is good. It would be even better if it put the number of electoral votes on each state, if it were accompanied by a table with the 2000 race percentages for each state, and if it could be frozen and saved after the user put in his own changes, but at least it is a start.

My forecast is Bush 328, Kerry 210, with the distribution of states in the map accompanying this entry.

The polls I trust most are at the Rasmussen Report. Note, by the way, that I, Eric Rasmus en, am unrelated to the Rasmussen of that report. I don't trust any polls very much, though. Bush still has a lot of ad spending to do, and the debates are still to come. Kerry will have even more ad spending (because of the Demo millionaires funding the 527s, where he's had something like a 60 million to 4 million dollar lead so far-- my unchecked guesses), and will also be in the debates. But so far the voters have been hit with huge amounts of free ads for Kerry in the form of the mainstream media's biased reporting. Bush's spending and the debates will go a little way towards evening that out.

Has anyone, by the way, calculated the value of the Democratic bias of the media? What I'd like to see is some estimate of the amount of newsprint and TV time that is effectively Democratic commercials, multiplied by the standard ad price for that time. I think it would swamp campaign spending, and would show that if we're really concerned about campaign finance reform, we should skip the small stuff- the billion dollars or so spent on paid-for ads-- and go right to the big money, which is the free ads in the form of news and op-eds.

Posted by erasmuse at 04:21 PM | Comments (0) | TrackBack