The other day, I noted that for all of the usual caveats about polls being snapshots and a week being a lifetime in politics, the May-June polls in 2004 correlated pretty well to the November results. A diarist at OpenLeft has posted a table showing the national presidential polling for March/ April, May, June and the actual vote margins for elections going back to 1972, along with the following observations:
- In the 9 elections since 1972, the leader at this point in the race has won a popular vote victory in barely half (5 times, 1972, 1976, 1984, 1996 and 2004).  Three of those elections involved a popular incumbent.
- In 7 of 9 elections since 1972, the final margin differed by more than 10 points from the May polling average
- In 7 of 9 elections since 1972, the Democrat has LOST ground to his Republican opponent from May to the final election.
- The last two elections have been the most stable from May to November since 1972.
Those last two observations may be the most significant. The current electoral maps do not look significantly different from the last two elections, suggesting another close race this year (absent the unforeseeable, natch). The general trend of Democratic nominees losing support by November gives John McCain hope of the sort Barack Obama is not selling.ÂÂ
OTOH, as I usually note, this year also seems to fall into a 16-year cycle of “change†elections, which has elected a Democrat narrowly in 1960, 1976 and 1992. Given the general mood of the country, Obama has a good chance at victory. It is still fair to say that neither candidate is favored at this point.
My problem with polls is less about what May-June means for November than what May-June means for May-June. Epistemologically speaking, I guess.
Polls, schmolls. Even for May and June these polls are next to meaningless. I know peering at tea leaves and trying to predict the future is very popular and compelling for some people and political writers are fascinated with polls, but we’d all be a lot better off if you guys would just stop fixating on them.
very informative post, Karl.
btw off topic, remember sometime ago we were discussing apostasy, when you quoted some know-pretend people from NYT and Commentary(?)or Weekly Standard(?), I know you were neutral about that.
But turned out I was right.
Did you read this:http://www.nytimes.com/2008/06/01/opinion/01pubed.html?pagewanted=1&_r=1&partner=rssnyt&emc=rss
“Given the general mood of the country…”
Well, yeah, if the general mood of the country runs something like ” Y’know, boy do I feel stupid today. I think I’ll just put this here revolver to my head and pull the trigger for Obama. Yeah. That’ll fix it.”
[…] interesting stats about spring time polls presented by Jeff Goldstein who blogs Protein Wisdom  (LINK) In 7 of 9 elections since 1972, the Democrat has LOST ground to his Republican opponent from May to […]
Karl,
Far be it for me to throw rocks at your “change election” theory, but methinks there may be a hole or two in it, if you look at the elections you cite closely.
1960 – While this does not get talked about much, the fact is that the Democrats stole that election, thanks to the Daley machine in Chicago. Yes, the election was close, and had it not been, even the Dem crooks in IL would not have been able to pull it out for JFK. Even so, the general consensus seems to be that had not the ballot boxes been stuffed by the DemoDalyCo, Nixon would have won.
1976 – I suppose you could call this a “change election”, although I always felt that Carter won more because most people just didn’t care that much for Gerald Ford.
1992 – This one I really don’t see as a “change election”. Had it not been for Ross Perot, the Clintons would have been a footnote to a footnote in history. They only received 42% of the popular vote, and had Perot not decided to act as a spoiler for GHWB (whom Perot hated), which is the real reason he became a candidate to start with, you can bet that most of the people who voted for Perot would have gone for Bush instead, even if they had to hold their noses to do it, given the choice.
Of course, I could be just whistling by the graveyard, but I do hope you are wrong. As much as I don’t like McCain, just the idea of a not-so-closeted Marxist like Obamasama anywhere near the White House makes me break out in a cold sweat…
The population is older and with more aging females than males than it was since 1972, back then we all thought we would be forever young so it was easy to be stupid. Geez it took me 40 years and Islamic Jihadists flying jets into civilian buildings for me to recognize the ‘brain-dead liberalism’ of my youth so I am never going back to that misery.
That said; I’m not so sure the aging radical feminists are going to let go of what O! has done to them.
sashal,
I had not seen that; it’ss not quite as definitive as you suggest, but I’ll add a note to the original post. Good catch.
Cave Bear,
You raise some good ponts about the 16-year cycle, some not so good.
In 1960, there were irregularities in YX as well, which is a point in your favor. OTOH, the dynamic of a less-liked VP following a really popular GOP prez might otherwise souynd like 1988. so why was it close in the first place? The differences were in the perception of the economy, the feeling that Ike’s generation had played out, and a charismatic Dem nominee.
1976: I think that post-Vietnam ennui, a lackluster (though improving) economy and a divided GOP had a lot to do with why Ford didn’t close the gap (though he came remarkably close).
1992: Perot is always a confounding factor here. On your side of the ledger, Perot took from traditionally GOP demographics. On my side, he did so by playing to the whole reform, anti-DC-biz-as-usual vibe that you see in both candidates this year.
Obviously, the 16-year cycle is not carved in stone. It’s just a way of understanding the sort of factors — particularly generational politics — that may be influencing this cycle. Confounding factors this year abound. McCain is also a plausible “change” nominee. Obama’s image is coming down to Earth. There’s a possible Bradley factor. The real economy seems stronger than the perceived economy. Etc. The 16-year cycle is not law — like the pirate code, it’s more like guidelines.
CB,
Also, we cannot assume there won’t be voter fraud in certain states this year, can we?
Cave Bear, Karl, and probably everybody else in the freakin’ universe:
Every election is different and polls and such are only one window of the process. Bob Dole’s campaign sucked, but that doesn’t mean he couldn’t have beaten Mondale. Plus, even Kerry could have beaten Ford if he was old enough to run then. But once Kerry was older, I guess he could have beaten a Dukakis or the elder Bush, but only barely. Carter was an Unstoppable Prez Machine once upon a time, and the current holder of the Oval Office has certainly had his polls wander a bit away from their earlier trajectory. And if Dole had run a decent campaign, my aunt would have liberated Auschwitz or whatever the saying is this year.
Also, attributing all the Perot votes for one candidate or another is the kind of thing only a Hillary Clinton ’08 campaign should do. You lost, get over it already. Or put up a better candidate.
I think we can assume there will be records amount of fraud. We might see Milwaukee submit more ballots than the rest of the state combined.
Will ACORN be found to be involved in any of the supposed fraud to come? How could that be?
Karl,
Good points, all.
jon isn’t much on “nuance”, is he?
Geez, were you ever as direct and to the point to the Dims about the 2000 election? It’s been eight fucking years and they are still whining about that one.
In the case of the ’92 election, it helps if you look a little close at the numbers. Clinton only got 42% of the vote, which is actually pretty close to average for all the post WW II Dem candidates who lost bigtime over the years. Which means 58% of the vote went to someone else.
Are you actually so dense as to think that anywhere near most of the voters who went for Perot were disaffected Demos? So if Perot had not run (and he knew full well he didn’t have a snowball’s chance in hell of winning; this was about fragging the hated Bush the Elder), just where do you think most of his votes would have gone?
Gore was a pretty bad candidate (Clinton baggage minus the Clinton support) who lost to a better campaign. I got over it, it’s history, and some people need to get over 2000 like others need to get over 1992.
And no, I am not so dense as to say that most of Perot’s votes would have gone to Clinton. In fact, I would say that Perot’s votes would have otherwise been pretty much a split of votes for Clinton, Bush, and whatever hopeless third-party guys were running then. You put Clinton’s 42% up as a point of ridicule, but it seems to me that Bush got a smaller number. And that proved what? Oh yeah, he lost. Sorry for the dearth of nuance or “nuance” or whatever it is I am lacking, but I just wasn’t in a pandering mode or whatever it is you want from your sycophants. I’ll try harder next time.
Polls, polls, polls.
I don’t know about you guys, but I think I can already hear the sound of The Big “O”‘s knees snapping.
CRACK!
Oh, nuts! Is that a racist word? Do I need to apologize to someone?
#9 Karl – do not forget that John Kennedy ran to the right on foreign policy and military policy. He wasn’t about hippies and VW’s, he wasn’t a beatnik. Other than charisma, he would be closer to GWB and JSM than BHO.
Presidential elections occur every four years, and there are so many changes between the last election and the current election that (no matter the election year) a lot of caveats apply in any analysis (you did caveat). There were changes in the current situation between 1912 and 1916, and between 1916 and 1920. Analysis of presidential elections seems to be of historical value and trying to pinpoint a few broad ‘do’s and don’ts’ than anything else. It appears to be closer to military strategy than any thing else – analyze the forces you face, the terrain, logistics, and what you have available before deciding on where and when to fight.