AsterPro Library of Credible Astrology
4. Criticism of Statistical Research by Gauquelins


Page as of Oct. 14, 2001

                       Statistical Astrology Research
                          Sirman A. Celayir, 1989

Research by Gauquelins. In the late 1980s, Michael and Francois Gauquelin,
who have studied psychology and statistics at the Sorbonne University,
and Dr. Suitbert Ertel of Göttingen University publicized statistical studies
about astrology. They claimed their studies validated astrology. The news
spread like a wildfire in the United States. The Gauquelins were then
employed by a book publisher (ACS) of San Diego that also advertised itself,
under the name NCGR, as an astrology research organization. The Gauquelins
arrived with much publicity in San Diego, to enjoy immense acclaim. NCGR
began publishing a series of articles about their work.

I reviewed the studies in the Spring 1989 issue of the NCGR Journal
and decided that they were more in the realm of number-juggling than
science. On May 7, 1989, I wrote the critique below and sent copies to Ms.
Gauquelin, the newly appointed editor of NCGR, and 150 copies to major
astrologers in the USA and abroad. The critique was published by Ms.
Gauquelin in the next issue, to NCGR's chagrin. Soon after that, there was
less talk about statistical studies. Sometime in 1990, less than a year after
their arrival, the Gauquelins left the USA, though NCGR maintained a P.O.
Box number for Ms. Gauquelin, probably for the sake of appearances, as if
she were still the editor. The letter is self-explanatory. The response by
Professor Ertel is presented at the end, and mine to his.

Dear Ms. Gauquelin and Professor Ertel:

I read your critiques in the Spring issue of the NCGR Journal. I found them
interesting and very academic. However, I suggest that you are both wrong
for the following reasons.

   ■ Confusion about the subset versus universal set. What is the basis for
your joint presumption that a natal chart is all that is needed to predict a
person's future profession?  Along with "planets do not compel, they
impel," one of the most basic axioms of traditional astrology is recognition
of the existence of influences that mold a person exogenous (external) to
astrology. There are hereditary, environmental, parental, and
circumstantial factors to consider. The astrological blueprint in a natal
chart at best shows a subset of potentials which evolve within the confines
of and in conjunction with these exogenous factors. Your works treat
astrological influences as the only pertinent variables and dismiss the
exogenous variables under a "ceteris paribus" (everything else remains the
same) restriction. It contradicts reality and common sense to dismiss them.
If you do accept their existence philosophically, then where are they in
your analysis?
   To be more specific, if the offspring of two genius parents is also a genius
and later becomes a first-rate scientist, is this attributable to the natal
configurations or to the genetic material transferred from the parents,
upbringing and guidance, schooling, and other external factors? Indeed,
are these not more consequential than the natal pattern? If so, then where
are these external factors reflected in your studies? As it is, statistical
models based entirely on natal potentials can only flag out marginal
patterns and correlations, and probably yield inconclusive results. Indeed,
this is exactly what both of your models yield: inconclusive results. A
"debate" about such results seems much too premature. The debate should
be about the sampling technique and the samples you have used.

   ■ The first maxim of astrology. By equating the subset of natal variables
to the universal set of "all" variables (of genetic and other factors), your
approach also questions the first maxim of astrology: planets do not compel.
You are implying that the planets do compel, that athletes are athletes
because their natal chart compelled them to be. Indeed, you may be testing
the validity of this maxim, rather than what you are claiming to test. And
since the results are not conclusive, you have proven that the first maxim
is indeed correct.

   ■ Sampling errors. Michael Gauquelin states (Page 36) that he "divided
3,142 'Men of War' from the private to the general and came to the 'easy'
conclusion that there were no differences between the two categories,
'neither for Mars nor for Jupiter'."  If I understand correctly his claim, it
seems a contradiction of common sense to pool together Audie Murphy,
courageous soldier, with Douglas MacArthur, one of the greatest generals
ever, together because both were soldiers. Gen. MacArthur was much more,
a mover of the masses, a brilliant strategist, and a man of historical
proportions whose vision has paved the path for modern Japan. One can see
MacArthur in the company of Kemal Ataturk, Mao Tse-tung, Eisenhower, De
Gaulle, Tito, and Castro, but not Audie Murphy. This would be analogous to
pooling oranges, grapes, and melons as the same fruit because their liquid
content is high, never mind that they look, feel, and taste different.
   The fact that Mars and Jupiter did not object to the pooling, to me, is not
an argument that supports the pooling. Rather, it is a strong indication that
you are using two irrelevant variables as pooling criteria. While Mars and
Jupiter may have some significance within the subset of astrological
variables, they are clearly inadequate in the universal set of all variables
that contribute to the outcome. Why? Because they cannot flag out immense
real-life differences among people included in your samples, like Audie
Murphy and MacArthur. Since you are using them as the only criteria, they
allow you to pool people as different as apples, oranges, and melons in the
same sample. Moreover, pooling 5000 people because they are all athletes is
likely to be counter productive, it will lead to unending debates about
marginal and inconclusive results, and it makes independent confirmation
of results almost impossible. I will suggest a different approach later.

   ■ Mixed focus. Unlike athletes and some other groups with verifiable
achievements (i.e., 9.8 seconds for 100-meter dash for designating a
sprinter as first-rate), there are categories of people in which first-class
rating may be both subjective and/or a surrogate for a more important
variable. For example, first rate actors and actresses may not necessarily
have earned their rating through acting skill, but by the fame and/or luck
elements which may or may not be present in their charts. Astrologically
and practically, fame, luck, and true first rate achievement are different
phenomena. Otherwise it would follow that packaging and appearances are
always synonymous to contents and substance. So, unless the sample is
carefully clustered, you could be studying a mixture of true achievement,
fame, and luck elements all at once without being certain which. This too
would contribute to your marginal and uncertain results.

   ■ A remedy. The following is an illustration that addresses the sampling
problems for athletes. For example, start with a list of world champion
boxers and top two or three contenders. It is said that boxing is a way out
of the ghetto for many boxers. This exogenous factor is likely to apply to
most boxers. The remaining ingredients, some of which may be genetic
makeup, choice of a violent individual sport, body strength and
coordination, and self-discipline are common factors that may fit into a natal
pattern. It is an ideal sample for a small pilot study. The investigation can
then extend to a similar sample of other athletes. A second stage analysis
can test the viability of pooling boxers with tennis players, for example.
This approach will also address unusual situations in which, for example,
the twin brother of a first rate athlete is not athletically inclined. Ideally
we all would like to see test results that associate specific natal patterns
with particular types of persons. Unfortunately, even if statistical tests do
show high levels of significance, it is doubtful that there will ever be
exactly matching natal blueprints for two first-rate soldiers. Nor can
astrology postulate categorically that a person who shares the natal
configurations of a first-rate soldier "must" also be a soldier. I think we
agree on this from the outset. This would further confirm the point I made
in the beginning: astrology is not the only mold that decides one's
profession. Thus, astrology may be forever restricted in its predictive
capability.

2) Scientific investigations should not lose sight of the forest for the trees.
It took you several years to collect your samples, to run statistical tests,
and to report on the results. Analyzing the real-life profiles of the persons
included in the sample (to have some control over exogenous variables),
would have added considerable time to your effort. All this, for a relatively
simple problem of how to differentiate athletes from scientists from soldiers.
     There are an almost infinite number of similar inquiries in astrology.
Specifically, "What is likely to occur in a couple's life when Saturn enters
the 7th house of one partner and forms a conjunction to his/her natal Sun
and a 1st house square to the other partner's natal Moon." Obviously one
cannot study each and every one of these inquiries through statistical
models each utilizing 3000 plus people. Indeed, if someone wanted to do
away with astrology, the best way of achieving such an end would be not
by outlawing it but by demanding a statistical verification for each and
every nuance.
     Judging from the remaining contents of the Spring issue of the NCGR
Journal, it would appear that statistical tests are here to stay. Apparently
the level of frustration has reached such a point that everyone wants to be
in the bandwagon of this type of research, even if just to hang on. Richard
Nolle presumably wants a higher degree of accuracy in his predictions for
the KGB in 1990. The article "Astrolabe and Astrological Research" reads
more like a list of what the company sells than research, and mid-July is a
good time for a conference, rather vacation, at Matrix offices in Big Rapids,
Michigan. (In case you are not aware, these two entities are major sellers of
software.)

3) There is a substantial amount of preliminary work that statistical
investigations overlook. Astrology is presently inundated with at least a
dozen House systems, one tropical and several sidereal zodiacs, and several
coordinate systems. One of the foremost tasks on the astrology agenda
should be doing away with the accumulated clutter. Imagine the task of
running a sample of 3000 athletes, scientists, and soldiers through a
combination of three zodiacs, two coordinate systems, and six house
systems. Then imagine the debate that would follow.

4) There is an even bigger problem brewing in popular astrology.
Thousands of people worldwide are attracted to astrology every year based
upon the contents and implied promises of dubious and unproven
astrological techniques and interpretations. Some consulting astrologers
use these techniques to advise people who are facing problems and who
lend their trust to astrology. Key members of the astrological community
and astrology organizations owe some responsibility to these persons as
they do to credible astrology. This can be done by a review of some common
practices (e.g., Progressions, Solar & Lunar Returns, etc.) and by
suggesting guidelines. These objectives should be higher on the priority
list. They cannot be addressed appropriately if attention is ad infinitum on
the interpretation of statistical results. Furthermore, the credibility of
astrology would be enhanced much more if some astrologers could add a few
humble responses to their vocabulary, such as "I don't know," "I am not
sure," or "this event cannot be analyzed or explained by astrology."


Responses to my Letter

1)  May 15, 1989: From Ms. Gauquelin, , P.O. Box 4322, San Diego, CA  92104

"Thank you for your very interesting letter . . . I will be glad to publish a
large extract from it . . . from the beginning to" (end of Part 1). "But you
must be aware that, if you agree to this publication, I will have to send your
critiques concerning Prof. Ertel's and Michael Gauquelin's articles to these
authors, so that they may send their answers to your critiques. But I
consider also the other parts of your letter as interesting and points of
view clearly expressed. I am presently writing a new book, 'The Horoscope
Revisited.'  I could cite them in this book, with your permission.'"...Best
wishes to you as a fellow researcher . . ."

2)  July 5, 1989: from James Russell, 7 de Warrenne Rd., Lewes, East Sussex
BN7 1BP, UK. (In July 1997, I was notified that Mr. Russell has passed away.)

"Thank you for sending me the monograph attached to 'An Invitation to a
Conference and Debate.' I find I can agree with nearly all the opinions
expressed in the monograph . . ."

3)  Professor Ertel's response appeared on page 38 of the Fall 1989 issue of
the NCGR Journal. Totally ignoring the points I made, he states:
"Apparently Mr. Celayir would have preferred unambiguous results and
firm conclusions. What I would like him to share with professional
researchers is tolerance of ambiguity, readiness to leave questions open if
necessary, and patience on our road to certainty."

My response to Professor Ertel's response

If the foundation of a research study is solid, I am more than willing to be
patient with preliminary findings. I am not so generous if the foundation is
totally fragmented, and the results of the study are offered under profound
titles such as "Reversed Eminence Correlations," as if they were blueprints
from heaven. What you seem to be suggesting is that I keep quiet about
these results until you publish your study in a "ground-breaking" book
on astrology and sell lots of it. Forget it.