Wednesday, November 07, 2012

What has the 2012 Election Taught Us?

©2012 J. Lee Lehman

First, congratulations to all my friends and colleagues who dared to predict the presidential election. I congratulate everybody, because the first step is to try.

I also hesitate to congratulate only those who predicted correctly because this kind of prediction is difficult, and fraught with problems. It is in the spirit of humbleness that I would like to address some of the problems I see in this prediction process, as well as to herald some things that I think we as a community are beginning to get right.

In the category of right, I congratulate the Political Astrology Blog and Chris Brennan and Patrick Watson for their work through the last two elections in providing a resource for predictions made, and timings recorded. It is extremely useful that the information from one election is not just quietly disappearing before the next cycle: and the next. We can only learn from developing a historical database. It is also for this reason that I urge everyone who has posted predictions or discussions of this election to not take them down. I know that nobody really wants to keep remembering their mistakes, but there is still gold to be mined in understanding both the "correct" and "incorrect" analyses - because hardly anybody gave single reason predictions. I would claim some small credit for Kepler College in helping to initiate a more serious discussion of these matters.

Here are some ideas I would throw out as important.

There are two psychological factors that we have to keep in mind that have major bearing, not only on our predictions, but those of the media, and citizens at large, whether of the country in question, or not.

  • It is an extremely well documented observation that virtually everybody makes the mistake of believing that other people agree with them more than is actually true. I have observed and commented for some time that there has historically been a high correlation between who astrologers predict will win with whom that astrologer will or would vote for. If we understand this within the context of this psychological tendency, we understand the very real peril: that if we believe that the universe is ordered, and we are right, then of course the universe will work out according to our own beliefs. While not a fully conscious process, this represents a considerable danger in making predictions.
  • When, as today, we are engaging in the post-election discussion, the tendency when one is wrong is to find a factor in the charts that one examined that could be construed as going in the opposite way, pounce on that factor, and then explain one's wrong prediction as being a result of that. So far, this may be plausible. But the real error is then the all too human tendency to believe that, having found the magic bullet, that one's prediction is transformed into a correct prediction, for having been explained.
In fact, predicting the outcome of political elections means using astrology in a fundamentally different way than we normally do. Natal astrologers are simply not generally called upon to compare two charts and see which one would be victorious in a contest held upon a particular day.

This simple reality is further complicated by the fact that everybody had a roughly 50-50 chance of being right or wrong. Thus, we are confronted with the probability that some of the correct predictions for Obama were actually fundamentally flawed in analysis and right by chance, whereas some of the predictions for Romney were just slightly incomplete, but primarily correct. How does one tell the difference, especially since we only repeat this exercise at four year intervals?

For now, let me address some particular factors, in the hope that they can help the thinking process for future elections.
  • The prediction of two-party models is completely different than parliamentary models, because only two party elections resemble the warfare models of which I am fond. If there are multiple armies on a battlefield, each one is not fighting all others: they are already aligned as allies, which is not how multi-party systems work.
  • I think we have to come to agreement for the future that if the US system is that the actual victory occurs through the Electoral College, that winning the Electoral College is the measure of a correct prediction. Frankly, we don't have enough data for a model of when an election is split between popular and Electoral College. However, I have to admit that I am being dragged kicking and screaming into developing some respect for the Electoral College idea. Had a hurricane or earthquake actually occurred on the day of the election, thereby significantly lowering a populous state's voting; or had a state referendum significantly changed the voter turn-out in just that one state; then the popular vote might not actually be more "just" than the Electoral College.
  • This election was the first failed case since Al Morrison pointed out the theory: that people nominated under a void of course Moon do not win the election. Obama was, and he did. My comment on this is twofold: first, there are no single factor arguments that I know that will work absolutely all the time. The second point is that the void of course has, imho, been blown up way out of proportion. It is merely one of the ways that the Moon can be afflicted. We must study them all.
  • This leads into the next issue: is there any critical event prior to the election itself that determines the outcome of the election? We don't often discuss this matter of multiple events surrounding an event, but we have the time a person announced candidacy, the time the election campaign is established, each of the primaries, the date of the primary at which the candidate receives the votes needed to become the candidate, the time that this is declared by the media, the time the nominating convention begins, the time that the candidate is announced as winning the nomination, and probably several other dates I am forgetting. Are any of these fail-safe predictors of the results? I tend to view these matters in the spirit of electionals in a string of related issues like a relationship (where one can do charts for the time of meeting, the first date, the first sex, moving in together, marriage, etc.): the most important thing may be defensive charting: that none of these event can predict a success, but any of them could preclude one.
  • Horary: it's time to give it up. Sorry, but let's stop kidding ourselves. In a world of seven billion people, the idea that any one of us citizens has the special pipeline to the truth about this is delusional. This cannot work, because of two irreconcilable problems: that we cannot know when the question was asked for the first time (making all subsequent attempts bogus with random access results), and we aren't really any more special than any other citizen of the Earth. I was especially disheartened by reading an argument that perhaps one gets the right answer if one has greater virtue. No comment. But I think the fact that there are many astrologers (me among them) who adore horary doesn't make this an appropriate technique.
  • It may do well for Americans to study European parliamentary elections as a model for the primary season. We simply don't have a model for what to do with multiple candidates.
  • In a world of early voting, it's time to revise our models. Dixville Notch is now a lovely historical asterisk, when literally millions have voted before those intrepid few have stayed up for their election night party. We may see a decline of importance of election day charts as being predictive, because so much of the voting has already occurred.
I believe that successful work on elections will occur through the combination of mundane techniques, and natal work focused around learning the appropriate factors to compare between timed charts of candidates: a database which now stands at seventeen cases where we have good data for both candidates. This is slow, hard work.

But most of all, I encourage my colleagues to persevere, always using caution about applying hindsight instead of foresight.

9 comments:

Chris said...

Excellent points, all.

Jamie Partridge said...

Hi Lee, thanks for the well thought out and written advice for us, especially the ones like me on the wrong side in this game. I agree with everything here except on little bit.

You said "everybody had a roughly 50-50 chance of being right or wrong. Thus, we are confronted with the probability that some of the correct predictions for Obama were actually fundamentally flawed in analysis and right by chance, whereas some of the predictions for Romney were just slightly incomplete, but primarily correct."

The polls always had Obama in front, the betting odds over the last 6 months averaged at $1.30 for Obama and $3.20 for Romney. So I would call it more like a 70-30 chance that Obama was going to win, going back a year at least.

Given those observations, would you agree that it wasn't so much a coin toss, and if so, how would that affect your point that "some of the predictions for Romney were just slightly incomplete, but primarily correct."?

J. Lee Lehman said...

Jamie, you raise a very interesting question here. Speaking less as an astrologer, and more as a citizen with statistical training, I was struck by the fact that the polls that were being published by the press - and here I must say both liberal and conservative - were wildly different (with the exception of a few out-lyers like Nate Silver) than was seen on the numbers from betting sites. So I do agree that in 2012, the odds were not 50-50. In fact, depending on how you handle vice presidents who assume the presidency upon the death of a sitting president, the odds of an incumbent president who is running for reelection being reelected are 69%. This is exactly like home field advantage in sport, and very much like the 70-30 you queried. I followed the betting sites through the election cycle, and they never wavered on the greater probability of Obama winning, although the actual odds shifted constantly.

However, for our purposes, my comment stands, because I have also observed that, until you have some greater familiarity with a predictive model than one or two data points can give you, there is no guarantee that your model can distinguish between a landslide and a 0.1% margin The point is, I have seen some outlandish reasoning applied to predictions of either man, and I have seen some good reasoning applied to predicting Romney.

I think a major crux of the difficulty is that we have no traditional model for elections, and from there, everything has to be done by analogy and judgment of how important particular factors actually are. This is a tough job, because normally we consider factors qualitatively, and here, it's like counting votes for number of factors for one versus the other.

Jamie Partridge said...

Thanks for clarifying that. US politics is a bit of a minefield compared to what I'm used to in Australia with the Westminster system. I like how our previous experience helps in our astrological understanding. My favourite subject at uni was bio-statistical analysis. The textbook was a nightmare and no software back them. Think I got about 20% in the final exam but ended up with a distinction thanks to the good old bell curve.

laughingcat said...

Hi Lee - Robert Wilkinson from The Aquarius Papers here. Glad I found this through Jamie's link. Every 4 years I do several articles on the election. One involves "real world" factors, one involves the Dixville Notch chart, and one examines a horary posited at some point before the election using traditional analytical means.

I quoted something you said in 2008 about Dixville Notch and prior lunations in my 2012 "real world factors" article on the election, but couldn't find much more and ran out of time. I hope one of my astrologer fans will pick up the work on prior Lunations, since I believe you're on to something there.

As for polls - Besides being an astrologer for over 40 years, I was a (former) professional political analyst in the 80s and 90s, specifically trained by the first person who ever did poll analysis from the 40s on and is considered a legend in DC for his accuracy and methodology. While most polls are worthless for a variety of reasons, and to some extent the model must be adjusted to real world realities put into play since election 2002, I do believe there are ways to predict with great accuracy which candidate will win which state.

I've been consistently successful in calling Presidential and Senate races, not so much with House races since (astrologically) it would require casting hundreds of charts for when polls open in each district. In House races, polls are almost worthless since in the absence of a crime or redistricting, incumbency trumps almost all other factors almost all the time. Polls may show internal trends in a district, but they're pretty obvious from the start even without polls.

Nate wasn't the only one who called it. Ezra Kline also does good work. Again, this election seemed to imply that aggregating the polls with an eye to internal poll bias gave accurate results.

As for EC totals, they can be predicted with a great deal of accuracy, given the lock that each party has on certain voting demographics. That's why this could very well be the last election where the Republicans will get this many EV. Population trends would seem to indicate this is the beginning of the end of the Republican party as a competitive entity with the platform it currently has.

Anyway, I was surprised to see your name pop up, and figured I'd come over here to check in, say hello, thank you for your past work, and introduce myself.

J. Lee Lehman said...

Robert - thank you! I had no idea that that was your background. I have not had the best of luck using the prior lunation from the election, but the Libra Ingress chart rocks.

I doubt we will have a final assessment of Florida for a bit yet (what else is new?), so we won't know exactly who gets the biggest kudos among the analysts.

This year, Anthony Louis attempted a state-by-state prediction based on state inceptions - the first time I think that an astrologer has done this.

I do agree with you on the demographic challenge to the Republican Party. The error in the Gallup Poll this year included a substantial underestimation of the Hispanic vote. As I discussed in Astrology of Sustainability, the USA is undergoing population trends which will not only have a long-term impact on the viability of the Republican Party as presently constituted, but the psychological impact of the sign change of Pluto into Capricorn will still be in a denial phase until 2020.

Kannon McAfee said...

Lee, I absolutely agree that personal bias too easily enters into astrologers' predictions and that this 'pick-the-winner' task is not the kind of thing that we normally do. However, that does not make it necessarily a difficult task astrologically, but probably a more difficult one personally because high interest level correlates to a strong political point of view. Eliminating bias and the wishful thinking it produces leaves the mind clear for better astrology. Balance and perspective - those are the keys to practice and study.

My approach to these predictions is actually pretty simple (or at least well-ordered). I use layers to ensure agreement of various factors.

First, what do the mundanes say? Is Venus a morning or evening star? This factor alone creates a 'favor' upon either the sitting leader or whoever the challenger will be. It is not always correct in determining the winner, but is a very good starting place.

Secondly, I look at the chart I use for the USA and see if 'change' is indicated by the transits. The most important historically that are easy to spot are any transits in *longitude & declination* to the South Node which benefits conservatives/GOP or to the North Node which benefits progressives/Democrats.

Thirdly, I do a thorough rectification and validation process on both candidates' charts (Presidential race) and make sure I am using a correct one that is validated by all previous life events. For many astrologers this is a brutal experience that they just don't want to undertake. However, I love the challenge and it has made me a better astrologer.

Ultimately, I am only concerned with knowing who will hold the Job on January 20th, not with how or how big they win.

Anonymous said...

Hi Lee, I disagree with your point about election day charts becoming obsolete due to early voting for a few reasons.
1) the vast majority of the vote still occurs on election day
2)I think the Dixville notch chart works out quite well this time (if you use the method for a sports match -which our election essentially is -rather than castle siege)
3)If the election day astrology isn't predictive, then why all the hullabaloo about Mercury stationing retrograde on election day (which Brennan and Watson were a part of)?

J. Lee Lehman said...

Gary -

I'm not giving up on Election Day charts - yet. But I'm raising the possibility that this may be the last, or the second to last, etc. You yourself have left an interesting comment in the ISAR E-Zine about the fact that the Mercury retro didn't do much this time. While I found your hypothesis interesting, and it may be true, you also made the case here that maybe the early voting is negating at least part of the effect of the Mercury retro on election day. There certainly were voting snafus, and hideously long lines - and it would have been a lot worse without early voting. My models actually didn't show anything related to the Mercury retrograde, so I'm not wrapped either way.