testing astrology with statistical methods

1
Statistical tests of astrological factors have failed to yield the kind of significant positive results that astrologers might expect, and those we might hope for. Initially, I think it seemed likely, if not certain, to those who see astrology in action that statistical methods would capture astrology?s truth. Although some tests have had significant positive results?notably the Gauquelin studies on the ?Mars effect??many other studies have failed or were not replicated.

Some astrologers take the view, for one reason or another, that statistical test results don?t much matter. They may not. Yet, as I see it, whether or not astrology needs to be confirmed with statistical methods, we do need to understand why it is that statistical tests have not yielded convincing positive results.

Astrologers have made various arguments to the effect that statistical methods cannot capture astrology. I have not found any of the arguments I?ve seen to be convincing. Suggestive. Possibly true. Part of the picture. But not convincing. In my view, however, there is a way to understand the problem astrology has with statistical methods, and the problem that statistical methods have with astrology. I find that way in the work of Jean-Baptiste Morin.

Morin, a 17th century French astrologer familiar to readers of Skyscript, was an impassioned reformer. He culled from the ruptured astrological tradition of which we are quasi-heirs what he thought was the true tradition of astrology. Morin may have thrown parts of the baby out with the bathwater. Yet, even if his some of his excisions from the tradition were excessive, Morin developed a wonderfully revealing theory of astrological signification and an integrated and fully articulated theory and method of astrological determination.

Morin distinguished a planet?s nature (its categorical meanings) and its celestial state (placement in the Sky) from its terrestrial state (placement within the astrological houses). In Morin?s theory, a planet?s nature and celestial state are inadequate to determine the planet?s meaning in a chart. According to Morin, the astrological houses determine planets? universal or general significations to earthly meanings and effects. In Morin?s theory and method, because a planet refers by nature to all its categorical meanings, it signifies by nature nothing in particular. Morin makes the strong statement that a planet signifies nothing?no particular thing?until its universal significations are determined in accord with its terrestrial state.

If Morin?s theory of astrological signification is applied to the design of statistical tests that attempt to test astrology, it becomes immediately apparent that, if Morin is right, statistical test designs are wrong. Statistical tests, with few exceptions, have tested astrology?s universal significators, often confining themselves to planetary nature alone. When tests come out showing that the planets did not mean what they were expected to mean, astrology is thought to be called into question. In truth, however, test results?both negative results and some positive results? tend to confirm Morin?s theory. .

As I see it, researchers might strike paydirt if they take Morin?s theory and method to heart. Morin set out an extensive system of principles and rules that enable effective application of his method. A test of Morin?s method would need to take into account all the principles and rules Morin specified and use them according to his theory and method. Those are quite a lot of rules and principles.

Statistical methods may or may not be able to adequately represent astrology for test purposes. It seems even more problematic to imagine that statistical tests can adequately represent what astrologers do. As I see it, however, unless statisticians already know that statistics are inadequate to represent a system as detailed and specific as Morin?s, we won?t know unless we try. It is important that those who seek to test astrology test astrology. It is important that we see that tests of astrological factors ripped from the system in which they function rest on an inadequate representation of astrology.

I?ve written at greater length on this topic in an essay published in Correlation: Seator, P. (2008/2009) ?Astrological Prediction and Statistical Tests,? Correlation, Vol. 26 (2), pp. 14-49. The essay is also published at http://encirclinglight.com/astrological ... ical-tests. One of the things the paper does is call into question and reject the analogy that Eysenck and Nias drew in their 1984 book, Astrology: Science or Superstition?, between astrological factors and factors that affect body weight. It appears to me that that analogy, or the assumptions on which it depends, underlies the mis-design of failed statistical tests of astrology.

Morin?s method and the astrological theory on which it rests may be a crucial key to understanding the relationship between astrology and statistical tests. The possibility deserves further investigation.

Thanks for this great forum, which I just joined, and for all the interesting posts.

penny seator
Last edited by penny seator on Wed Dec 02, 2009 5:56 pm, edited 1 time in total.

3
Hi, Ed,

Yes, I think so. There are other studies that rely on the angles, too, that have come out well. Kenneth Irving found significance for an angular Moon--and referred to the angles as "where the money is." There is the Hill and Thompson study on the Mars redhead effect.

Even the fact that Mars came out strongest in the Gauquelin studies I think is telling. In Morin's method, analogy activates shared significations of house and planet. Mars has strong analogies with the ascendant as ruler of Aries, the sign that sits in the ascendant in the archetypal Aries rising chart. Mars has strong analogies also with the Midheaven. The midheaven, as the point of Sun's culmination, is the hot point and Mars is hot. Morin uses exaltation rulers of houses. Mars is the exaltation ruler of Capricorn, which falls in the 10th in the archetypal Aries rising chart.

I really think there's something here and would love to see researchers investigate the possibility. I also think Morin can help astrolgers work out questions of astrological theory and method.

Thanks for your comment.

penny seator
Last edited by penny seator on Thu Dec 03, 2009 12:35 pm, edited 1 time in total.

4
I think there is some value in using statistics in astrology, just like they have their value in psychology and other social life issues. However there seems to be a paradox in astrology. On the one side astrologers reject statistics because it is said that astrology is too complex to be caught in statistics. On the other side astrology works with certain rules. The statisticians would oppose to this that these rules could never be known if they can't be tested.

Perhaps astrology is indeed too complex or too flexible to test, but this makes that a strict judicial astrology would be very difficult to defend. As I feel the most attracted to 'natural astrology', I don't have problems with this. When I see squares in a chart I can say that the native will have some problems to deal with in life somewhat on the level of the involved planets, but I can't exactly tell what these problems are. Same with prognostication, an upcoming square transit of Saturn may indicate some troubles in ones mood but what actually will happen is hard to tell.

I don't feel the need to explain everything with astrology and can accept that astrology is just an aspect in life amongst social circumstances, political situation in a country, family life, free will, or just plain coincidence and of course the traditional view of 'divine intervention'.

This side makes astrology also a (part of) social science; to make a good interpretation of a chart it will be necessary to see the native and know a bit of his/her life and circumstances, then astrology can fill in some gaps and some timing. This makes astrology prone to psychological effects like 'cold reading' and hindsight interpretation. So be it. It is good to realize this though; knowing how the interaction between astrologer and native is affected by such effects should not be rejected or denied I believe, but serves a better understanding of these effects and therefore of astrology which then can be distinguished from them.

5
I've posted a little about this topic on threads before but in retrospect I don't think I explained myself fully enough - my comments were interpreted as just supporting some people's view that astrology does not lend itself to statistics. In fact I was stating something more fundamental and I would like to explain it again here with some examples that illustrate my point. It will be a bit long but I think that is essential to fully explain the issue.

Statistics, in the way it is being discussed here, is essentially inference based baysian probabilities. The key assumption of such statsitics is that a sample tends to be representative of a population in a certain mathematical manner. That I think everyone gets.

( population in this regard is any type of population - group of people, insects, houses, whatever but I will stick to groups of people for my illustrations)

The key to the statistical theory working is, that it assumes that every sample drawn from a population is likely to be equally representative of the population - and, in addition, as the number of samples drawn approaches infinity the profile of all the samples together will approch the profile of the population. Well obviously, you say, but the mathematical theorty that stats depends on requires both statements to be true adn as we will see one maybe flawed ( not mathematically but in application to our real universe)

Don't see the problem yet? Read the first requirement again. What it is saying is that, if I draw a sample now ( 1 a.m on 5th Jan 2010) from the list of everyone in London called Anne (to investogate let's say something non-astrological;the incidence of dog ownership), that sample will be as equally representative as if I draw a sample, via thesame randowm number generator, from the same population (and ignoring deaths and migration) at 15 hours on 7th March 2012. It also says that if I draw a sample now and Joe draws one now they will also be equally representative.

Now do you see the problem? Astrology quite categorically says that each moment is unique. Indeed horary and electional astrology depend on it. My sample drawn now will differ in its representativeness from the one in March 2012 because the universe's unique planetary configuration at those times will make it so. Similarly Joe's sample according to astrology will not be equally reprentative of the population as mine - because Joe has a different natal chart etc and so a sample drawn by Joe will reflect Joe as well as the population. Maybe Joe has more affinity with dogs according to his natal chart...

Oh dear. See where that leads us? It leads us, if we believe in astrology, to the conclusion, not that one cannot test astrology with statistics, but that one cannot infer anything about anything with statistics.Astrology undermines one of the key assumptions on which statistical inference of anything is based.

Now, I hasten to add that this does not mean that a population does not have a given profile. If one looks at 100% of a population one will get the exact number of people called Anne with a dog ( which if you want to know that particular bit of information is fine). I can conclusively tell you how many Chairmen of FTSE 100 Companies was a Pisces in 2006 becuase I looked at all 100 companies. It is only the use of sampling that doesn't work. But its application fails not just for astrology, as many argue, but for inferences made for all tests.

Of course this is scary - prescription drug safety for example depends on statistical inference. But that is not the point of my post. The point is merely to put the debate about demonstrating astrology through statistics into the correct framework.

6
amelia wrote:Statistics, in the way it is being discussed here, is essentially inference based baysian probabilities. The key assumption of such statsitics is that a sample tends to be representative of a population in a certain mathematical manner. That I think everyone gets.

The key to the statistical theory working is, that it assumes that every sample drawn from a population is likely to be equally representative of the population -

Astrology quite categorically says that each moment is unique. Indeed horary and electional astrology depend on it. My sample drawn now will differ in its representativeness from the one in March 2012 because the universe's unique planetary configuration at those times will make it so. Similarly Joe's sample according to astrology will not be equally reprentative of the population as mine - because Joe has a different natal chart etc and so a sample drawn by Joe will reflect Joe as well as the population. Maybe Joe has more affinity with dogs according to his natal chart...

Oh dear. See where that leads us? It leads us, if we believe in astrology, to the conclusion, not that one cannot test astrology with statistics, but that one cannot infer anything about anything with statistics.


Hi, Amelia,

Thanks for the very clear statement of your argument, and that you explained it as a problem of sampling. I?ve heard the basic argument about the unique nature of each moment, but without the problem cast clearly as a function of sampling biases.

This argument, even even in your version that highlights sampling biases, doesn't resolve the problem that astrology and statistics have with each other. To the extent that the goal is to create some kind of bridge between the astrology and statistics camps, or points of view.

You incorporate this into your statement. ?It leads us, if we believe in astrology, to the conclusion?.? It works for astrologers to see this possible source of sampling bias, but unless you have already seen astrology in action and know what you see, the argument from sampling bias is a bootstrap argument. It says that statistical methods cannot show whether astrology works because astrology works. If you don?t see that astrology works?and if you?re in the statistics camp you mostly don?t?then you?re left entirely unmoved. You shrug your shoulders and say, ?Yes, that would be true if astrology were true, but it?s not.? So it can't move the discussion--if the purpose of the discussion is to resolve something about this apparent impasse.

This argument may be correct. But it seems to me that the problem that practitioners of the traditional art can see, and that Morin's theory and method make clear, is the more fundamental and incontrovertible one: To test astrology, we need to test astrology.

Thanks for your clear statement, and for explaining the problem with samples.

Penny
Last edited by penny seator on Sun Jan 17, 2010 7:17 am, edited 1 time in total.

7
penny seator wrote:5. In horary, a question needs to be sincere and important to the querent. Other rules about whether the chart is radical need to be applied. Drawing a statistical sample is outside the rules of the game. (The next point is related to this one.)
I think horary could be tested statistically. The horary astrologer could collect data and write down every time what the answer was and later check for it with the client. The danger is if one doesn't register the results, one might only remember the positive results of enthusiastic reactions like a wallet found on the exact place as mentioned in the horary. If there are answers that were wrong it could be more likely that the client doesn't call back to say that the answer was wrong.

I hardly know anything about horary but it seems to me that with yes/no questions like "Am I pregnant?" it would be easy to check this with a personal statistics. In this interview with Maurice McCann he tells how he tries to be sure to find the answer by finding the best method. So quite a rational way.
6. We enter a mantic field when we divine. We?re in a conversation. Our interlocutor is willing and able to take in different circumstances and intentions and enter into what are in effect stipulated definitions of the project at hand. Maybe cosmos--or the daemon we're working with--participates appropriately in the project at hand.
I think there is some truth in this but on the other side this wouldn't support the case of astrology. Maurice McCann's answer to such a question shows his view.
The interview with Maurice McCann wrote:Q: The thing I wonder with cases like that - does the faith, the confidence, in the mind of the astrologer have an impact on the results do you think?


No, I think it's sheer hard slog. It's not a case of, if you're in a magical mood, it'll all work for you; it won't. I'm telling you it won't! What you've got to do is, you've got to know all the components of the chart. You can't overlook a rule, or the fact that, three weeks later, there is Mercury - it's coming up to conjunct Mars, but before it gets there it goes retrograde. (..........)
And it is very mechanical. The planets are going around up in the sky in a mechanical order. They just go on, boringly doing that. Saturn is boundaries, structure, and those guys (the planets) are doing the structure all the time - until they explode or something.
His view seems to make it ideal for testing.
But Rudolph Smit performed a test similar to that and thought it put astrology in doubt. As I recall he tested to see whether in a sample of cases of accidental death a secondary progression of Mars to the ascendant or of the ascendant to Mars was in effect. His comments are on his website, www.astrology-and-science.com, or were a couple of years ago in a piece on Mr. Smit?s personal journey with astrology. He considered his finding that these progressions were not significantly correlated with death by accident to be evidence against astrology.
www.astrology-and-science.com (or www.rudolfhsmit.nl) is a good website (it was the website I referred to in the 'dark matter' thread). Although 'non believers' these guys are ok, not the typical 'debunkers' but rather neutral, having neither interest for astrology being true or not. During the holidays I had no disposition of the internet. Previously I had printed every article on the website and have read them all.

I must say that this could have quite an effect on my dealing with astrology. I had seen the researches in Dean and Mather's 'Recent Advances', and while houses and signs gave negative results aspects were given a chance, and the prospects were optimistic. Since I follow Kepler's view (emphasis on aspects and not using signs and houses), the negative results for signs and houses didn't bother me.

Much of their optimism had waned (according to the astrology-and-science people), many of those tests hadn't been performed rigidly. A test performed by Peter Niehenke, an astrologer (off all people), on aspects in natal charts turned out to be negative. http://www.rudolfhsmit.nl/d-rese2.htm ?4) So aspects aren't spared either, so for me this was bad news.

The last few years from while to while I've been feeling dissatisfied by 'personal' (natal charts, and mainly my own) astrology observing transits etc. I notice that I too have retreated more in the 'subjective' view of astrology (see the explantation on this in the 'dark matter' thread).

I therefore think of researching weather astrology (following Kepler's views) for a while. Quite concrete to research and 'non-personal'/mundane of character. I feel quite optimistic after some positive results. ( http://skyscript.co.uk/forums/viewtopic.php?t=4263 it turned out to be the best spring in years ; and http://skyscript.co.uk/forums/viewtopic.php?t=4922
two weeks ago there was extreme snowfall and cold). However I realize that a more structured research is required.

Koen van de Moortel is a Belgian astrologer (also mentioned on Rudolf Smit's website) who also does research www.astrovdm.com
Interesting is his 'Love at first sight research' http://www.astrovdm.com/attract.htm

This illustrates that a dialogue between astrologers and (scientific) researchers can be possible, and that's a positive thing.

8
His [Maurice McCann?s] view seems to make it ideal for testing.
I'm not so sure about that Eddy. Maurice McCann has been a good friend and a generous colleague to me, even though we have very different perspectives on some aspects of our shared interest in horary astrology. I have never been to understand his objective perspective, expecting horary to perform like a science regardless of the querent?s need to know the question asked. The last time I spoke to Maurice, about a year ago, he wanted it to be publicly known that he had more or less dropped out of astrology. After years of studying horary statistically, and asking questions in order to study the results of his charts, he found it was unreliable. This comes as no surprise to me; nor does it shake my conviction that horary is a very powerful and reliable powerful divinational tool, which knows how to stay ahead of our efforts to profit from it in anything other than a sincere, heartfelt, query.

Although I didn?t always agree 100% with Maurice, I leant a lot because of his influence. He was critical of anyone claiming to be a horary astrologer who didn?t know the traditional techniques thoroughly, or who quoted popular authors rather than traditional texts, so he kept us all on our toes. But ultimately I think the lesson is this: don?t expect to understand horary and use it effectively, if you think you can manipulise the system and bend it to your own will. It does not respond to scientific probing or personal ego, and anyone who feels in need of anyone else's approval is probably not ideally suited to being a horary astrologer.

I hope 2010 is a good year for you Eddy,
Deb

9
Relative to Ms Houlding's post mentioning Maurice McCann's years of horary "experiments"-the results of which led Mr. McCann to the conclusion that horary is "unreliable"-I wonder:

-how many of his horaries were undertaken when the Moon was VC? I do not believe Moon VC negates horaries but-shall I state it stongly-I KNOW that Moon VC definitely tends to render horary indications unreliable.

-in how many of Mr. McCann's horaries was the Dragon's Head posited in the 1st House? 13th Century Gerard of Cremona' declares without reservation against the reliability of any horary with such a placement.

-in how many of his horaries were Lots (Parts) or dodekatemorions used in the delineations? Both (especially Parts) were considered by the Arabs-originators of horary-of much importance in horary delineations.

-in how many of Mr. McCann's horaries were fixed stars in close longitudinal conjunction or parallel of declination with planetary significators, sensitive degree points (like the ascendant and MC), Arabic Parts (and the dodekatemorions of all these)-fully (or even partially) taken into account in the horary delineation? Fixed stars-according to tradition-can (in close conjunction or parallel) potentially radically influence any chart and may significantly modify the prediction/answer to the horary question.

When experimenting with something we need to clearly and precisely define our experiment; if our experiment is to test results relating to an entire field, such as the field of horary ( in Mr. McCann's case the field of "traditional" horary) then such an experiment must incorporate all of the elements and historical considerations involved, not merely a selected number of those elements and considerations.

10
Moon is VOC about 10% of the time. Node would be in H1 about 12% of the time rendering about 22% horarys unreadable. That's still 78% readable. Even with added lots/parts restrictions surely there would enought % charts that are readble - or are there too many restrictions that render horary essentaily useless? Maybe Carr has a point :???:


Martin

11
I think both Deb and Dr Farr have brought up excellent points.

Since we're still in the realm of subjective here (no scientific verification), in my experience, and probably in at least quite a few other astrologers' experiences, horaries about trivial questions simply do not work. I have no idea how you'd quantify that in a science experiment sense - how could one define what's trivial? And wouldn't trying to do that take you right out of the scientific territory you're trying to ground this in?

The second problem, the one that Dr Farr talked about, is what constitutes a qualified astrologer? Some of us are competent at it but even then not everyone is competent in every field of astrology (you may be a wizard, at say, daytrading charts, and know nothing about decumbiture), all of us are still learning, and sadly, astrology being a method of divination, you're going to find lots of totally unqualified people, as well as lots of fakes.

The only things I can think of are - like Eddy said - maybe weather prediction, or possibly something like sports astrology per Bonatti's rules of warfare, as Bernadette Brady and Lee Lehman have been doing (and keeping statistics on). I don't know a lot about sport, but it does have the advantage of win/lose/draw/game cancelled as its only outcomes, Bonattis's rules are spelt out pretty clearly, and the result is easily verifiable, as are Bonatti's rules. What kind of extra-astrological factors you'd have to weight for, I know not. But it seems like something like that might be small enough to be tested, and big enough to still be real astrology. Of course it wouldn't prove whether all astrology worked, it would at best prove that that bit of astrology worked or did not work when you had presumably competent astrologers doing it.

It's incredibly difficult to narrow astrology down enough to make it scientifically testable (at least if you want the astrology to be something more than sun-signs or the like), but I do think you have to look at pretty small pieces at a time. And how to make sure (somehow, but how could we know?) that the astrologers involved are competent with the astrology involved? An astrological peer review process seems difficult, as most well-known astrologers are known for various types of modern astrology, not traditional. It wouldn't hurt if the scientists understood the astrology, either, but it's not something you can pick up overnight.

12
Martin Lewicki wrote:Moon is VOC about 10% of the time. Node would be in H1 about 12% of the time rendering about 22% horarys unreadable. That's still 78% readable. Even with added lots/parts restrictions surely there would enought % charts that are readble - or are there too many restrictions that render horary essentaily useless? Maybe Carr has a point :???:


Martin
No, there are not too many restrictions that render horary useless, but there are considerations (frequently overlooked) that must be taken into account in order to undertake the horary question/delineation with any hope for accuracy.

My questions relating to McCann's results address whether or not in his horaries he took the above strictures (and other factors mentioned) into consideration. By ignoring the above, could his results have been rendered unreliable? You mention 78% chart readability after deducting for probablity of Moon VC and Dragons Tail in 1st. Suppose (not considering VC and the node strictures) McCann achieved a 60% accuracy rate in that 78% of horaries free from VC and DT1st: however, if McCann did not take into account that 22% deduction for strictures, and had based his statistics on 100% of the horaries, he would believe that he had only gotten a 47% accuracy rate (less than a 50/50 or chance result) Certainly he would consider such a finding indicative of unreliability!

The Lots/Parts (dodekatemorions and the fixed stars) are not "restrictions" in any sense: at one time (prior to about 1400) these considerations constituted a major element in horary. They are important elements to consider when making a horary delineation. However, they have played little role in (Western) horary over the past several hundred years: in evaluating horary charts, isn't it possible that many could get "wrong answers" due to ignoring these anciently-important elements in the delineation? I suggest that McCann's results were disappointing because his experiment (probably) did not incorporate all of the considerations and elements required for accuracy in horary technique.