The facts, ma’am, just the facts

by Daniel on September 19, 2004

By way of a break from everything about the US elections in the blogosphere, here’s a post about the US elections.

Swift Boats blah, forged memos blah, Kerry campaign blah blah. The latest thing I’m hearing is that some polls are biased against Kerry (I’ve linked to a comparatively sane version of the thesis here; the other kind is not exactly hard to find). Forget about all that, give me something I can work with. One of the most important lessons that experiences teaches you when you start hustling economic data for a living is the importance of picking an indicator and sticking to it. If you choose a number that you regard as representative and look at its evolution over time, you’ll end up getting a feel for the reality that it represents. If, on the other hand, you start chopping and changing the way you look at things, you will, with probability 1, confuse yourself.

The CT fave indicator is the Approval-Disapproval Spread Point Chart. Henry linked to Nasi Lemak’s version to illustrate a point about optical econometrics. I used something similar as the basis of my abortive trading system on the Iowa Electronic Markets. Paul Krugman linked to it once but I can’t be bothered chasing up the link. It’s a good indicator. It was Lee Atwater’s number of choice, and it captures something important about the public perceptions which underly voting behaviour. Sure, it isn’t a direct question about voting intentions, and it doesn’t capture some of the horse race flavour of the headline polls, but I think it would be a mistake to dump it now.

(There will now be a pause while everyone clicks over to Dr. Pollkatz to look at the chart)

Have a look. The really ropey period for Kerry in the polls has coincided with a flattening out and perhaps even a small increase in the Bush A/D spread. It almost looks like this line is having one of its periodic bounces – one might label this one “Republican Convention”, the other two being “Iraq War” and “Saddam Captured”

Anything to do with the Swift Boats would not affect people’s perceptions of George Bush. Similarly, biases in pollsters’ treatment of registered voters for the two parties ought not to show up in the A/D spread figures where the same adjustments aren’t made. Therefore, I would advance the argument that these two phenomena are explanations which cannot explain the whole dataset, and should therefore be discarded in favour of theories which can.

In fact, looking at the data, I am attracted by the following four bald assertions:

1) The recent Bush rally is both within the normal variability of the data and not out of line with past rallies. It just feels so much more important because it’s closer to the election date.

2) The rally is, however, genuine. It’s not likely to be an artifact of polling, and it’s too spread out in time to be linked to any particular blogospheric news story of interest.

3) It is notable that the Bush A/D spread has flattened out more or less exactly when it reached the zero line. This is unlikely to be a coincidence.

4) The temptation is very strong, as noted above, to label the most recent spike upward the “Republican Convention” effect.

I can’t prove any of these, so to some extent I am talking out my arse, and am contributing to the quadrennial pile of election-related bullshit. However, all four of them are at least consistent with the data, which I hope puts them in the top quintile of that pile.



anon. 09.19.04 at 10:50 pm

The CT fave indicator is the Approval-Disapproval Spread Point Chart.

But then you assume that the election is only a referendum about Bush, not a contest between Bush and Kerry.

I think one can summarize the current electoral climate in the US the following way:

(1) Voters are not really convinced that Bush is doing a good job, thus the divided A/D poll.

(2) But voters are also not convinced that Kerry is a better candidate than Bush, thus Bush’s by now considerable lead in various head-to-head polls.

In short, Kerry needs to show the American voter that he’s more than ABB (Anybody But Bush).


Giles 09.19.04 at 10:52 pm

A second thing to note is that previous highs have come about through disoncitous jumps followed by the gradual depletion of the gain – so there’s a spike when Saddam is captured followed by dissipation as people come to realize that this one event doesn’t change things much.

The recent upturn looks much less like a discontinuity and perhaps signals a trend. Specifically it may suggest people are drawing positive conclusions on the basis more than one piece of evidence. This suggest that the electorate are starting to act more rationally and less emotionally. Secondly there is then the question of whether decisions made on the basis of more than one piece of evidence are more persistent – i.e. will there be less tail off from the current minor spike.


Zizka 09.19.04 at 11:16 pm

Giles has an uncanny ability to draw conclusions where others are unable to do so.


KCinDC 09.19.04 at 11:34 pm

What do you think the multiple pieces of positive, rational evidence might be, Giles?


Giles 09.20.04 at 12:08 am

I’m not going to get into whether their rational but I think rathergate has put a positive spin on Bush’s air service, economic indicators while not stellar are positive and Iraq seems to be huming along just fine – or more particularly, not as bad as everyone was predicting 4 months ago.


roger 09.20.04 at 12:26 am

Going with James Surowiecki’s The Wisdom of Crowds thesis for a moment, I think you’ve picked the wrong indicator. The right indicator, to me, is in the question; who do you think is going to win? That is the indicator that gives us the kind of information that sets the odds on boxing matches and horse races. And it is way up for Bush — 60 percent in most polls. In April, according to the Times, it was 49 for Bush.
That’s a pretty bad trend for Kerry.


Ralph 09.20.04 at 12:43 am

Iraq humming along? Not as bad as everyone was predicting?

Is somebody broadcasting from a different planet?


Zizka 09.20.04 at 1:03 am

Ralph: I’ve been finding ever-increasing evidence for the many-worlds theory during the last few months. Philip Roth’s new novel about President Lindbergh is as realistic as a lot of stuff I hear people saying.


John Isbell 09.20.04 at 1:23 am

Based on the empirical data that the Earth is flat, for which there is non-trivial evidence, I can reach all kinds of remarkable conclusions. I must share them with you sometime, it will rock your world.


epist 09.20.04 at 1:39 am

My hasty and uneducated perusal of blog writing on the polls convinces me of a couple of things: 1) that there are two sets of data sources, those that have Bush up by more than 10, and those who have him up by less than 3. 2) The EC map is looking wierder, but Kerry is more likely to pull ahead in PA and NH, since he has polled well there all year, than Bush is to pull out CO, NV, ME, and WV, all of which just crawled into his favor in the margin of error.

So my ex rectum argument is that, given this data, the bounce has to get better for Bush for him to win.


llld 09.20.04 at 2:16 am

Interesting post. Regarding the last comment, I think it is unlikely that Kerry will win a close election. The Republican party seems to be much better positioned to use unethical practices or cheat. They run a ruthless operation.

Both parties can and have used cheating (illegal) and unethical practices in American elections. But the more ruthless operators at this juncture of history seem to be the Republicans. The unethical tricks they have pulled off so far are numerous, including
i) staging a “union” guy assaulting a man with his daughter.
ii) pushing through a phony convicts list in Florida
iii) getting an operative to force an investigation of Kerry’s medals (talk about craven)

I am sure many readers can list more. So when the voting machine results come back in Ohio and Florida, I expect them to closely reflect Bush favorable polls. In fact, an uncanny resemblance is very likely.


frankly0 09.20.04 at 2:30 am

Of course, the very problem with the latest polls, which problem you simply dismiss, has a clear impact on Bush’s approval ratings.

If the number of Republicans is considerably overrepresented in these latest polls — and all kinds of evidence strongly suggest that it is — then Bush’s approval numbers would be unrealistically high.

It would be a very useful thing to create a similar graph corrected for plausible Republican/Democrat representation. My guess is that mostly the shape would be the same, with the only difference being that the final Bush bounce would be much diminished, and already in rapid decline.

And it is hardly reasonable to regard the overrepresentation of Republicans as just something to which we should pay no attention. It is an authentic anomaly, producing wildly, and otherwise inexplicably divergent polls. This is not a problem that goes away by ignoring it.


frankly0 09.20.04 at 2:35 am

The recent upturn looks much less like a discontinuity and perhaps signals a trend.

Huh? Does your graph go on longer than mine??


dnexon 09.20.04 at 2:36 am

If polls remain roughly the same by election day, I predict with total confidence that George Bush or John Kerry will be elected President.

A less snarky comment: if the state polls look like they do now by election day, the key factor will be turnout. We’re talking about within MOE spreads or barely outside of MOE spreads in enough states to swing the election. I repeat, if little changes, we really won’t know.

Why were the polling predictions wrong in 1998, 2000 and 2002? Differential turnout. And we simply don’t have a good way to predict that.

I spit on the wisdom of crowd :-). Those numbers only matter if they lead to a significant combination of bandwagoning (people like to vote for winners) and depressed turnout for the group that doesn’t expect to win.


Zizka 09.20.04 at 2:41 am

Bad polls may have the beneficial effect of making Bush overconfident if he believes them. For example, he apparently thinks that WV is safe and MN winnable. If he mistakenly moves resources from WV to MN, he hurts himself.

I think that the primary effect of these polls is to discourage Kerry people, though.

I’ve toyed with conspiracy theories of the big (10-point) discrepancies but am not committed to one.


James Surowiecki 09.20.04 at 4:00 am

“I spit on the wisdom of the crowd.”

Well, that’s the first time I’ve heard that one. Makes all the other reviews seem positively glowing.


racer 09.20.04 at 4:27 am

Let’s hear ’em, Zizka.

If there’s money to be made by corrupting polls, I’m 95% confident that it’s being conspired about as we blog.


evans 09.20.04 at 5:52 am

Is it just me not understanding the chart, or aren’t Bush’s approval/disapproval numbers lower than when he took office, even lower than August 2001, and probably on a continued downward swing? So that would be a nice spike up on a path down to his personal low?


rc 09.20.04 at 6:28 am

Like Prof. Pollkatz, I analyze the job approval spread here; but in addition, I also look at the trial heat spread. If you look, you’ll see that I use the two together to calculate poll bias or lean for each factor.


evans 09.20.04 at 7:09 am

Is it just me not understanding the chart, or aren’t Bush’s approval/disapproval numbers lower than when he took office, even lower than August 2001, and probably on a continued downward swing? So that would be a nice spike up on a path down to his personal low?


Ragout 09.20.04 at 5:15 pm


The criticism mostly isn’t about “biases in pollsters’ treatment of registered voters for the two parties.” It’s about the fact that the polls have shown a large jump in Republican ID. The critics would like to adjust this away.

Since it’s shown up in multiple polls, the spike in Republican ID clearly isn’t due to the bias of particular polls with bad methodology or due to sampling on particular days.

It’s a genuine Bush bounce, presumably due to the convention and the 9/11 anniversary. There are also some signs that the bounce is dissipating already (e.g. the recent Pew poll).

I’ve been saying this for weeks. The bounce is genuine and the poll critics are letting their partisanship override their better judgement:


BigMacAttack 09.20.04 at 5:22 pm

A good way of detecting trends but how useful is it for predicting?

How bad does Bush’s A/D ratings need to be before the US will elect a Democrat? Apparently pretty bad.

I would be really worried about this if I were a Democrat.

I think the debates and the spin from the debates will be important.

But right now I am amazed at how well Bush is doing. (In part this is probably because I find Bush so under whelming)

I live in NJ and I actually saw some Elect Bush and Elect Kerry commercials last night. That was topped off with a front page story in my non-liberal local newspaper with the headline Thinking the Unthinkable about a possible Bush victory in NJ. That is really bad news for Kerry.


sagenz 09.20.04 at 10:00 pm

so, uh what happened with rather. try for some new info


brkily 09.22.04 at 4:25 am

if polls are being manipulated to show bush ahead, it would certainly make bush wins election results – less suspicious and harder to challenge…

Comments on this entry are closed.