The Race is Already Over

I recently attended a lecture by Allan Lichtman (see here: http://en.wikipedia.org/wiki/Allan_Licht man). Among other things he is a prognosticator of presidential elections and has developed a system "13 keys to the white house" that has correctly predicted the popular vote winner in every election since 1984. I believe his system also retrospectively predicted the winner in every election since 1860.

(edit: corrected the keys)

During the lecture Lichtman explained his 13 keys--it's a binary "true/false" system. What was fascinating is his system shows that things like experience, media bias, and campaigns in general make no difference in presidential elections. In fact, most of the keys can be answered well in advance of the election year.

Here are the keys:

KEY 1
Party Mandate    After the midterm elections, the incumbent party holds more seats in the U.S. House than it did after the previous midterm elections.
KEY 2
Contest    There is no serious contest for the incumbent-party nomination.
KEY 3
Incumbency    The incumbent-party candidate is the sitting president.
KEY 4
Third party    There is no significant third-party or independent campaign.
KEY 5
Short-term economy    The economy is not in recession during the election campaign.
KEY 6
Long-term economy    Real per-capita economic growth during the term equals or exceeds mean growth during the previous two terms.
KEY 7
Policy change    The incumbent administration effects major changes in national policy.
KEY 8
Social unrest    There is no sustained social unrest during the term.
KEY 9
Scandal    The incumbent administration is untainted by major scandal.
KEY 10
Foreign/military failure    The incumbent administration suffers no major failure in foreign or military affairs.
KEY 11
Foreign/military success    The incumbent administration achieves a major success in foreign or military affairs.
KEY 12
Incumbent charisma    The incumbent-party candidate is charismatic or a national hero.
KEY 13
Challenger charisma    The challenging-party candidate is not charismatic or a national hero.

Six or more "False" means the challenging party will win, in this case Obama and the Democrats. The questions have some subjectivity to them and therefore must be answered free of partisan bias. According to Lichtman, there are EIGHT False Keys this year, therefore Obama will win.

HOWEVER, there is a major caveat this year according to Lichtman: Race. Race is not factored in to his keys and the effect of race in the election is difficult to quantify. But all else being equal, this race is over and has been over for about a year:)

Tags: Alan Lichtman, obama, predictions (all tags)

Comments

23 Comments

Re: The Race is Already Over

Very interesting.  I have seen variants of this before, but not this specific list.

Recc'd

by AntiStipes 2008-09-06 12:32PM | 0 recs
Re: The Race is Already Over

this is fascinating, but the way the questions are written the whole "8 false" thing makes no sense.  I only count three false answers.  
could it be the way you transcribed them?  Because otherwise this makes alot of sense.

I believe, at the end of the day before election day, all the smears and scandals don't amount to much -- folks vote for whoever they think will either change their lives for the better or maintain the status quo.  if the status quo is kicking their ass?  They'll vote for change and all the pitbull attacks and smear campaigns won't make a bit of difference.

Take 2004 -- the status quo for most folks was they hadn't lived through a second terrorist attack and nothing else was beating up the majority all that much at the time.  

Now we have economic problems reaching into the vast majority of households if not in real terms in job insecurity, healthcare costs and falling home values.  

With "hero" and "charismatic" cancelling each other out it will all come down to the economy and, sadly, race.

But on the race issue, we should have expected as much going in.  If we allow that mentality to shape our choices we will never break the ceilings on race and gender.  

And that said?  I think age and race could cancel one another out in this election.

by GRO 2008-09-06 01:00PM | 0 recs
Re: The Race is Already Over
Bigdaddy was a bit sloppy. Some of the criteria he listed as positives when they are meant to be negatives. Check here
KEY 1
Party Mandate    After the midterm elections, the incumbent party holds more seats in the U.S. House than it did after the previous midterm elections.
KEY 2
Contest    There is no serious contest for the incumbent-party nomination.
KEY 3
Incumbency    The incumbent-party candidate is the sitting president.
KEY 4
Third party    There is no significant third-party or independent campaign.
KEY 5
Short-term economy    The economy is not in recession during the election campaign.
KEY 6
Long-term economy    Real per-capita economic growth during the term equals or exceeds mean growth during the previous two terms.
KEY 7
Policy change    The incumbent administration effects major changes in national policy.
KEY 8
Social unrest    There is no sustained social unrest during the term.
KEY 9
Scandal>    The incumbent administration is untainted by major scandal.
KEY 10
Foreign/military failure    The incumbent administration suffers no major failure in foreign or military affairs.
KEY 11
Foreign/military success    The incumbent administration achieves a major success in foreign or military affairs.
KEY 12
Incumbent charisma    The incumbent-party candidate is charismatic or a national hero.
KEY 13
Challenger charisma    The challenging-party candidate is not charismatic or a national hero.
by vcalzone 2008-09-06 01:08PM | 0 recs
Re: The Race is Already Over

Updated diary to correct the keys. thanks.

by bigdaddy 2008-09-06 01:57PM | 0 recs
Re: The Race is Already Over

No problem. I knew what you probably meant when I first read it.

by vcalzone 2008-09-06 02:20PM | 0 recs
Re: The Race is Already Over

well hold on now.  With the new key I get 10 false answers, not eight, which is that much the better. I only get 4, 7 and 12 as true with 11 up for debate.  

Whatever.  I love this way of looking at it.

by GRO 2008-09-06 07:28PM | 0 recs
Re: The Race is Already Over

Sadly, during a summer in which I worked for a professor of mine as an assistant, he confidently touted a similar formula as predicting the victory of Al Gore.  

Not to be a downer, but these things change over time due to electoral shifts or something as small as a shift in the electorate in one state.  Think about Missouri.  Is it REALLY a bellwether anymore?  Or just the red version of Pennsylvania...a close state that often goes one direction over the other?  My guess is that, in cases where Dems win big, Missouri goes blue, but generally red.  This is all off the cuff, so I'm sure there are problems with it, but you see my point.  Things change.  But if Obama wins and does a good job, race will matter much less (among the swing voters and working class) in the future.  

by freedom78 2008-09-06 01:05PM | 0 recs
This is superstition

Not science.  The basic concept is sound, certainly--find a system that fits existing data and then test it against new observations to see if it can correctly predict.

However, we have had so very few Presidential elections that anything based on them is not statistically significant.  That's why political pollsters ask hundreds or thousands of people for their opinions, not just a few dozen.

And the elections since 1984 is an even smaller sampling--there have only been 6 elections in that timeframe.

Look at it this way--if I just grabbed a coin and flipped it, I could correctly predict all the elections since 1984 with a 1 in 64 chance.  In other words, if I got myself and 63 of my closest friends together, and we all flipped coins and compared our results to the elections since 1984, odds are at least one of us will match the data.

It doesn't mean that person has a magic coin that can predict elections.

Nate over at fivethirtyeight has a pretty good explanation why political superstitions are useless.

by BishopRook 2008-09-06 01:28PM | 0 recs
Re: This is superstition

The way I was thinking of it was this: find someone who has voted for the eventual winner ever election since 1984.  You could present him as having predicted the last six elections.

Except that there are probabably close to a million people that that describes.

by MeganLocke 2008-09-06 01:45PM | 0 recs
Re: This is superstition

"Not science."

Except that it is:

http://www.pnas.org/content/78/11/7230.a bstract?sid=6e701af1-fe9e-4a56-bb18-cdcc d1795181

Pattern recognition applied to presidential elections in the United States, 1860-1980: Role of integral social, economic, and political traits

  1. A. J. Lichtman* and
   2. V. I. Keilis-Borok†

Abstract

The outcome of American presidential elections in 1860-1980 follows certain regular patterns which can be described phenomenologically by simple integral parameters of "common sense" type.

by bigdaddy 2008-09-06 02:01PM | 0 recs
Re: This is superstition

Quoting from that paper:

In the statistical test of our conclusions we reject the simplest competing hypothesis--that the outcome of an election is independent of our dignosis of the situation.  We neither claim that other parameters cannot be used for the same purpose nor suggest methods for predicting future elections.

Their paper shows that they've created a model that can account for the 31 elections between 1860 and 1980 more accurately than random chance, but which has no predictive value and may or may not be better or worse than any other similar political model.  Bravo.  That's a very low bar to have set for themselves.

I'll concede: it's very lazy science, but it's science.

Quoting from an article by Lichtman in 2007 (emphasis mine):

The election for president is more than a year away. Neither major party has as yet chosen a nominee. Yet the results of the 2008 election are already in: the Democrats will recapture the White House next fall

...

This good news for Democrats and grim news for Republicans comes from the "Keys to the White House," a historically based prediction system that I developed in 1981, in collaboration with Volodia Keilis-Borok, an authority on the mathematics of prediction models.

...

The keys, however, are a robust system that has endured through momentous changes in the electorate, the economy, the society, and the technology of elections. It is unlikely that any contingency will alter the negative verdict on the party in power.

That is not science.  That is overselling the conclusions of your paper.  And in fact, it's suggesting that their predictions are foolproof, which is beyond the realm of science to prove.  Science can only say "this model fits the data we know of now."  Science cannot say "it is unlikely that there will ever be new data that disproves this model."  Any "scientist" who tells you that is trying to sell you something.

by BishopRook 2008-09-07 08:49PM | 0 recs
Re: This is superstition

Oh, and I should further add, when Lichtman says this:

a robust system that has endured through momentous changes in the electorate, the economy, the society, and the technology of elections

...I have a feeling the "momentous changes in the electorate, the economy, the society, and the technology of elections" he's referring to are not what have occurred since 1980.  He's referring to what's occurred since 1860, which is the earliest data point he examined when making his model.

That's some clever sleight of hand, because it isn't true.  In order to say that your model has endured since 1860, you have to show that the model you would have created if you only had 1860's data would survive the test of time until today.

As a matter of fact, it's likely that if he looked only at 1860 to 1920, say, then there would have been some deficiency located between 1920 and 1980 that broke his model, but that he was able to smooth out by looking at the full 1860 to 1980 range.

He's saying it's "endured" those 120 years because it verifies the exact same 31 elections that he used to build the model.  That's simply not how science is done.

In computer science, for example, when you want to test out the effectiveness of a pattern recognizer, you take a set of known data, and split it up into a training set and a verification set; then you train the model on the training set, and test on the verification set.  You never, ever, ever test on the training set, because that tells you nothing about predictive value towards unknown data points.

My professors wouldn't accept that in an undergrad term paper, nevermind a published work.  Again, not science.

by BishopRook 2008-09-07 09:08PM | 0 recs
Re: Gore, Kerry- Not Too Faste

Do you know how many prognosticator of presidential elections confidently predicted a solid win by Al Gore in 2000 ? Virtually every single one of them.

Are you aware that the overwhelming majority of these so called prognosticator of presidential elections predicted a sure win by John Kerry in 2004 ?

After a perfect batting average for many election cycles , these professors have been proven wrong in the last two elections.

Remember that these last 10 years have been the most politically & culturally divided country that we'd had.

Conservatives vs. Liberals
Republicans vs. Democrats.

Party loyalty, Race, Gender, etc.

A divided line that never really existed full blown until now. There were no blue regions vs. Red regions'

Yes, any Democrat running for President is suppose to be as heavy favorite in this anti-GOP national mood.

But the way things are looking now, this will be a nail bitter until the very end.

Plus these so called prognosticator of presidential elections are dealing with something major with no precedence or history ever.

A black man as candidate with a name Barack Obama.

Nor have they ever dealt with any possibility of the " bradley effect" on election day.

Whether right or wrong, Not a Single Person really knows if Obama's race will be a major factor on election day. Not a single person really knows if a "bradley Effect" will actually happen or not.

by latinomaker 2008-09-06 01:42PM | 0 recs
Re: Gore, Kerry- Not Too Faste

Except Litchman....

He had Gore winning the popular vote  in 2000 and W winning the 2004 election. I'm not talking about "all prognosticators" I'm talking about Allan Lichtman.

One confounding factor he talked about was the micro-effects that a 3rd party candidate can have. Nader polled less than 10% in 2000, but it made a difference in Florida. Moreover, he found that in Florida, 1 out of every 50 white-area-votes was tossed, but in black-precincts, 1 out of 9 was tossed. That amounted to around 50,000 votes tossed from predominantly AA neighborhoods.

by bigdaddy 2008-09-06 01:51PM | 0 recs
Re: The Race is Already Over

Sorry, I pulled the keys from another site. I don't have the handout Lichtman gave me with the keys--I left them at work.

It was 8 False and here are some notes from Lichtman:

-McCain is not considered a national war hero. Only Ike and Grant had that stature.

-Bush has not been tainted by personal scandal during this term (whether we think so or not)

-There is no serious 3rd party challenger. You need to have someone polling at 10% for this key to turn.

-There was no serious fight for the Republican nomination. Fights among the challengers are irrelevant.

-Bush has no major national legislative achievements this term.

-I believe that he considered there to be no civil unrest and that we're not in a recession--but I'd have to check my notes.

I guess having the keys and answers in front of me would have been nice before starting this diary, huh? LOL.

The most interesting thing according to his system, IMO, is that Swiftboating, Willie Horton (?), Experience, Obama vs. HRC, etc, didn't/don't matter.

by bigdaddy 2008-09-06 01:47PM | 0 recs
Re: The Race is Already Over

These things are amusing, but ultimately useless.  Every presidential election is different.  EVERY ONE.  Someone may find a set of keys that predicts a handful of elections, but there is no guarantee it will correctly predict the next one.

For a long time people called Missouri a bellwether state, saying no president had been elected in modern times without carrying it.  Well, it's pretty obvious that Obama could easily win the presidency without that state this year.

by Skaje 2008-09-06 02:14PM | 0 recs
Re: The Race is Already Over

Well I think this system is rooted in science and the scientific method, and that its been nearly flawless in its predictions. However, given that race is a new variable in this election, you have to take the prediction with a grain of salt. Bottom line is these predictions don't matter anyway, since the race still needs to be run and we'll have a "REAL" winner (not a predicted one) soon enough.

by bigdaddy 2008-09-06 02:28PM | 0 recs
Re: The Race is Already Over

This race has been close and will remain so right up until election day, in my opinion.  In the end though, I think Obama will come out on top mostly because his message has been so incredibly consistent over the past year.  

by Will Graham 2008-09-06 02:22PM | 0 recs
Re: The Race is Already Over
What is this? Everything's negative day on the liberal blogosphere?
Can't you little old lefties in tennis shoes have a positive response to anything?
by spirowasright 2008-09-06 02:24PM | 0 recs
It's only negative if you're pulling for

the GOP.

I saw the person in question explain the metrics, and it does bear an intrinsic logic.  These are key bellweathers of the national mood as it pertains to the incumbent and challenging parties.

Like all systems, events could occur that would contradict its predictions, but all things being equal it holds together.

All it really says is that it is a Dem year, and that history bears out these trends.  Don't stop working for it.

-chris

by chrisblask 2008-09-06 04:14PM | 0 recs
Re: The Race is Already Over

From my poll watching activities, I notice no major shifts from the 2000 and 2004 elections. Basically, the election comes down to Ohio and Florida again. Obama's got a better chance in Iowa, Colorado, Nevada, New Mexico, and Virginia than last time, McCains got a better chance in New Hampshire, Michigan and Pennslyvania.

My guess, its going down to the wire in Ohio. The election is coming down to a few thousand vote shift in Ohio.

by Zzyzzy 2008-09-06 04:40PM | 0 recs
which Keys are false?

I count 1 and 13 as unambiguously false, what are the remaining six?

Lichtman got my hopes up in 2000, and this is about the right time for him to show up. He gets debunked every cycle.

This looks like another close election, who wins depends on who works hardest and who doesn't do stupid things like underestimate the competition, or assume the race is won.

by souvarine 2008-09-06 05:08PM | 0 recs
Re: which Keys are false?

RE: 2000

...and Gore won the popular vote. What's the problem? His system predicts pop vote winner.

by bigdaddy 2008-09-07 08:44AM | 0 recs

Diaries

Advertise Blogads