New Statistics on the Gauquelin Keyword Study
Zip Dobyns
Our last Newsletter mentioned the return of a publication devoted to research in astrology: Correlation, published by the Astrological Association of England. I received a copy of the first issue soon after the Newsletter went out and was immensely impressed and pleased with it. There are several excellent articles in the issue. If future issues maintain the level of this initial one, it will make a real contribution to astrology and should be read by all serious astrology students. Among the bits of news included was a report on a panel presented at the spring conference of the British Psychological Association which included papers on astrology! Our own American Psychological Association has yet to present astrology in any form, as far as I know. In fact, I just attended the annual conference in late Aug. 1981, for the first time in several years, and had made tentative plans to try to present a program for the 1982 conference. I have presented astrology programs for the Association for Humanistic Psychology at several conferences, going back to 1968, but the A.P.A. is THE voice of psychology in the U.S., and as far as I know, astrology has yet to breach its walls.
To return to the article to which I referred in the title, Michael Startup is the author, and his work involved a statistical analysis of a study carried out by Francoise Gauquelin comparing the keywords published by 14 astrologers back to and including Ptolemy with the descriptive words in biographies of famous people. The discussion of Michel Gauquelin’s recent book has mentioned the association of planets in Gauquelin’s “zones of high intensity” with certain character traits. Thousands of adjectives were taken from biographies and certain words were used about people with planets in the zones (basically the zones were the mutable houses) with high odds against the results occurring by chance. It was noted that the words associated with people who had Mars in the zones of high intensity were very similar to the description of Mars in many astrology books. The words describing people with Saturn in those zones sounded very much like astrologers’ descriptions of Saturn. The Moon and Jupiter were the other two planets for which the Gauquelins had a list of descriptive adjectives with high statistical odds against chance: that is, people described with those words had the particular planet in one of the mutable houses more often than would occur by chance.
Francoise took the descriptive words of 14 astrologers for those four planets and compared them with the words used to describe people. In Francoise Gauquelin’s statistical evaluation of the results, each of the keywords used was assumed to have a 50% chance of being a hit or a miss. That is, a word used by the astrologer to describe the Moon might appear on the list of words used to describe people who had the Moon in a mutable house, or it might not be included in the group of words for people with the Moon in those “zones of high intensity.” Using this approach, Gauquelin found little significance in the astrologers’ lists.
Michael Startup claims that the chance expectancy is not 50% but more like 20%, since there are five possibilities. The astrologers’ words may apply to any of the four planets or to none, comparing them to the subjects whose biographical descriptions were the source of the project. The percentage is not exactly 20% since it depends on how many words are included in the group of words for each planet, and this varies from planet to planet. Startup used the chi square to test the significance of the planetary words, and found that even the ancient astrologers performed around 50% while some modern astrologers were well up into the 60s and 70s. One astrologer, Henry Weingarten, achieved 85% but since his book was written after the Gauquelin’s research was published, it is probable that he was familiar with their findings. When Startup converted his results to a chi square, he was not able to find tables going high enough to actually figure the odds against chance! A chi square of 32.8 gives odds of 100 million to one, and the astrologers’ odds included 4 over 100 and 2 over 200.
Startup performed a further statistical calculation to go beyond simple percent of coherence, which might occur with disagreements as well as with agreement, and to go beyond chance agreement; a statistic called Kappa developed by Cohen. The accuracy of the astrologers (in agreeing with the words associated with Gauquelin’s empirical findings) put Weingarten in first place (questionable as noted above because of prior knowledge of the Gauquelin’s work), Barbault of France in second place, Margaret Hone third, Paul Grell fourth, and Jeff Mayo fifth. The outcome can be considered either support for high reliability or for high validity, depending on whether you consider the Gauquelins’ work valid in some final sense.
Startup ends his article with: “Previous research by the Gauquelins has shown that at least 5 of the planets have distinct effects, related to their positions in the diurnal circle at birth. But now we have evidence that astrologers have a good idea of what these effects are and have been in possession of this knowledge since at least the 2nd century A.D. We are now faced with the dual tasks of explaining not only how the planets come to have the influences that have been discovered but also how astrologers come to have knowledge of those influences.” p. 42
Note the unquestioned and unacknowledged premises in the preceding quotation. Influences are assumed. Use of the word “effects” in the context is another way of assuming that the planets somehow produce the result—the human personality. What will it require to get people to look at their implicit materialistic assumptions? At least one writer for the journal being discussed is aware of the philosophical issues. The lead off article in this first issue of Correlation is by Patrick Curry, and he contrasts the symbolic vs. the physicalistic view of astrology. Curry describes approaches of a number of philosophers of science, ending with a short description of a recent group and their “methodology of scientific research programmes.” This MSRP approach takes Kuhn into account but offers a way to avoid getting stuck in an attitude of “all models are arbitrary.” The MSRP approach does seem to offer a practical system for dealing with the human obsession with model building. I still have a sneaking suspicion that all models are to a large extent arbitrary and that all we can ask of a model is that it be useful. It is possible that many different models could be useful in dealing with a complex reality. Maybe there is no one “right” model.
I do recommend that anyone who is serious about astrology and science should subscribe to the journal. Members of the Astrological Association pay 3 pounds for surface mail. Non-members pay 4.5 pounds. Your bank can give you a money order for British sterling. Checks go to Mrs. F. Griffiths, 98 Hayes Road, Bromley, Kent BR2 9AB, England.