The Gauquelin work
The universe is fantastic, don't you realize that?
Abstract -- Until his untimely death in May 1991, Michel Gauquelin (b.1928) was the world's most formidable scientific researcher into astrology, where his studies rank among the best ever conducted. Much of his success was due to Francoise (1929=2007), his Swiss-born wife and co-worker until 1985. His early fascination with astrology led to 45 years of research that resulted in a dozen popular books (many translated into several languages), 30 data books, and about 150 scientific articles. He used rigorous methods and large samples of hundreds or thousands of cases. Tests of signs, aspects, transits, and astrologers were negative. Tests of planetary positions with respect to the horizon were positive but only for eminent professionals and only for the Moon, Venus, Mars, Jupiter, and Saturn. The effect sizes were tiny (typically around 0.04 when expressed as a correlation) and of no practical use. Tests of planetary links between parents and children seemed positive but the effect size was even tinier (barely 0.02) and later seemed to disappear, leaving their existence in doubt. The planetary effect was later called the Mars effect because Mars (the planet linked to sports champions) was then the focus of attention. But depending on the occupation (there were nine others) it could have been called the Moon, Venus, Jupiter or Saturn effect. There was no effect for the Sun, Mercury, Uranus, Neptune, or Pluto, or for ordinary people such as those who visit astrologers. Planetary effects among eminent professionals have so far replicated across a total of 34 out of 35 studies, of which 8 are by independent researchers. Gauquelin's negative results upset astrologers, who found them hard to explain, and his positive results upset skeptics, who found them equally hard to explain. The outcome was 45 years of attack from both sides and a legacy of baffling puzzles for both astrology and science. Includes the Gauquelins' own account of how it all started, a non-technical account of his approach and results including negative results, a bibliography of main works, and nine historic photographs. Part 2 (a separate article) takes up the crucial topics of artifacts and how the baffling puzzles might be explained.
The following historical account of the work and findings of the Gauquelins is based on their many books and articles, and on 15 years of personal correspondence and meetings. Even if you are familiar with their work you may find much of interest. Here "Gauquelin" means Michel Gauquelin (1928-1991) and "the Gauquelins" means Michel and Francoise, his Swiss-born wife (1929-2007) and co-worker until 1985. Appendices give their accounts of how it all started, his experimental approach, and his early negative findings. Part 2 (a separate article) takes up the crucial topics of artifacts and how the baffling puzzles revealed by the Gauquelin work might be explained. For tributes to Michel and Francoise see elsewhere on this website.
But were the claims of astrology true? He was unconvinced by the success of his readings and by the opinions of astrologers. However, during 1919-1946 the European astrologers Paul Choisnard, Karl Krafft, and Leon Lasson had separately published statistical surveys with apparently positive results. Could he believe them? To find out, Gauquelin started collecting and testing birth data, an activity that would continue for the rest of his life and change the face of astrological research.
In 1946 he enrolled at the Sorbonne (famous seat of the faculties of science and letters at the University of Paris) to study psychology and statistics, graduating three years later, during which time he continued testing the Choisnard-Krafft-Lasson claims. In 1952 he met Marie-Francoise Schneider, a Swiss psychology student whom he later married in 1954, and with her encouragement he wrote his first book. It was a highly productive partnership that lasted for more than thirty years until their separation in 1982 and divorce in 1985; she continued with her own research after his death, and died in 2007 after a decade of declining health. Gauquelin describes these early days (it was 1948-49) as follows:
"At the age of 20, I was wildly enthusiastic about everything to do with astrology, although I was almost equally mad about painting and tennis. I wanted to do research into astrology, but had neither the money, nor the time (I was finishing my studies), nor the application necessary. I did, however, have a foggy idea of what was involved. I would go down to the library, between a drawing session in a Montparnasse attic and a tennis match, and copy out the names and dates of birth of people ... It was a very long-term, expensive and obstacle-strewn project" (Truth about Astrology pp.19-20).
After a year or two of working like this, he had a few thousand cases without birth times, and a few hundred cases with birth times obtained from registry offices. See Appendix 1 and 2 for descriptions by Michel and Francoise of how it all started.
State of astrology in the 1940s
In other words astrology in those days lacked adequate birth data and methods of analysis. Gauquelin's idea was to provide both, which (with the later help of Francoise) he did to an extent never matched before or since. The Sorbonne had taught them the skills needed to achieve standards well beyond those of previous researchers such as Choisnard, Lasson (France); Krafft, Schwab, Von Klockler (Germany); Carter, Sunley (UK); Bradley, Church of Light, Tobey (USA). Eventually their data (many tens of thousands of cases all with birth certificates), methods, and results filled thirty large data books, a dozen popular books many translated into several languages, and about 150 scientific articles and monographs.
Today Gauquelin's studies rank among the best ever conducted in astrology. Nevertheless his popular books tend to ignore astrological research by others (eg most of those cited above) and thus give an incomplete view of the field. But this was a common fault among early researchers, largely due to a scattered literature and the absence of reviews, an absence not remedied until the arrival of Recent Advances in 1977. See Appendix 3 for more details of Gauquelin's approach to filling the above double void.
The book received good coverage in the French press. For example there was nearly a full page in the popular weekly France Dimanche (30 Dec 1955:20) headed "Trois planetes decident de votre profession", which included a small picture of a young Gauquelin holding a page of figures with the caption (in French) "During ten years, the young French statistician Michel Gauquelin has drawn up millions of figures to prove astrology false. But today he concludes that certain planets exert an influence." There was a similar large article in the only daily sports paper L'Equipe (The Team, 21 Dec 1955:5,9) headed "Faut-il croire a l'influence des astres?". Later there was a long article in the popular monthly science magazine Science & Vie (January 1959:87-91) headed "Astres, Destinee, et Mathematiques" above a large picture of a slide rule resting on the famous zodiac lady from Les Tres Riches Heures du Duc de Berry, with a subsequent smaller picture of "M.Gauquelin, mathematicien du destin" at his desk. So Gauquelin's book was fairly well publicised, at least in France.
Gauquelin's conclusions were generally bad news for astrology and were as follows:
Choisnard (1867-1930) was ignorant of statistics and generally used rather small samples (one or two hundred cases). The results given in his La Loi d'heredite astrale (1919) and Preaves et Bases de l'astrologie scientifique (1921) did not exceed chance fluctuations and did not replicate with large samples. Choisnard was also ignorant of astronomical effects, for example he attributes great significance to superior people having more rising signs near Virgo-Libra than near Pisces-Aries, but this is simply because (in the northern hemisphere) the former rise more slowly than the latter due to the varying angle between horizon and ecliptic.
Despite these disappointments, Gauquelin acknowledges that Choisnard inspired his application of statistics to astrology, for which Choisnard's tables of planetary positions were "easy to use and were for a long time indispensable to my work". Choisnard had also inspired his use of birth data taken from registry offices. "Choisnard was, without doubt, the first to have had the idea of using this valuable source of information bequeathed to us by the Committee of Public Welfare" (Neo-Astrology 1991:158).
Krafft (1900-1945) believed he knew the answers in advance of any research. His Traite d'Astro-biologie (1939) "forms the most indecipherable mass that one could imagine ... He never gives the raw data on which his arguments are based. He replaces them with a debauchery of 'suggestive' graphs [that] have the malicious gift of making statistics show what they do not show at all" (L'Influence des Astres 1955:39). Like Choisnard's, Krafft's results did not exceed chance fluctuations and did not replicate with large samples. His genius for error was extraordinary, for example in a sample of 115 musicians he attributed great significance to the absence of Moon conjunct Uranus in certain signs even though the slowly-moving Uranus could not have been in those signs in the first place, see Figure 1.
Gauquelin comments "It seems quite incredible that he could have made such a ridiculous error. However, such errors occur on almost every page of Krafft's book [which] is a kind of statistical anti-methodology [that] should be given to students of probability calculations to put them on their guard" (Astrology and Science 1970:135).
The claims disconfirmed by Gauquelin ranged from simple ones like zodiac signs vs personality to more complex ones like planetary aspects between family members and transits at death. For details see Appendix 4. Gauquelin comments that his findings represent "a considerable inquiry in the testing of astrological rules with large and varied samples. It is necessary to stress that the results demolish astrology more than they might appear ... because they attack not the claims of particular authors but the elementary basis of the doctrine itself" (L'Influence des Astres 1955:62).
It was also a methodological challenge. For statistical purposes Gauquelin divided the diurnal circle into 36 sectors in order of diurnal motion starting at the ascendant, ie in the opposite direction to houses, see next section. However, the expected frequency in each sector is difficult to calculate because it depends on season (longer days in summer mean more births above the horizon), demography (the birth rate varies during both the day and the year), astronomy (planetary motion can be highly non-uniform eg due to retrogradation), and a particular sample (for all of the previous reasons). Gauquelin managed to solve this problem in various ways that were later published as Methodes pour etudier la Repartition des Astres dans le Mouvement Diurne (Paris 1957, 112 pages), which included a section by Francoise on three ways of calculating sector positions (rise-set, Placidus, graphical).
Gauquelin noted that Lasson's sample sizes were inadequate (often less than a hundred cases), and that he had not allowed for astronomic and demographic factors. For example Lasson found that writers tended to be born with Mercury rising, which in fact is due to the larger number of births near sunrise and to Mercury never being far from conjunct the Sun, so astrology has nothing to do with it. In the end Gauquelin was unable to replicate Lasson's results.
Nevertheless, much to Gauquelin's surprise, he observed in his own large samples similar (but not identical) results that did replicate. And it is these results, not the failed results of Choisnard, Krafft, and Lasson, that fill most of Gauquelin's first book. In its introduction he notes that he was "sincerely persuaded that [his book] would have nothing to offer except a critique of astrological doctrine ... But in the course of our work we were confronted by results so remarkable that scientific rigour obliged us to pursue and extend them" (L'Influence des Astres 1955:12). Furthermore "no a priori reason appears sufficiently valid for rejecting the conclusions we have drawn" (p.15). He had experienced astrological observations that, in the famous words of Johannes Kepler (1571-1630), "had instructed and compelled my unwilling belief."
Note that the above "we" and "our" refer (in the style of French writing) to Michel only and not to both Gauquelins. For the first few years after they met in 1952, Francoise gave mainly moral help and was not involved in the writing of his first two books, even refusing an invitation to participate in the second because she was studying for exams. Planetary effects were discovered by Michel alone, a point missed by many commentators, for example by Suzel Fuzeau-Braesch in L'Astrologie (1989) in the vast "que sais-je?" series by Presses Universitaires de France, which error was corrected in the 1992 edition.
Replication leads to two general hypotheses
Each profession was linked to one or more of three planets, namely Mars after which the effect is named, Jupiter, and Saturn (the Moon was added later and Venus later still), but not to the other planets or the Sun. For example the planet showing a surplus was Mars for physicians, military leaders and sports champions, Jupiter for actors, and Saturn for scientists. Mars and Saturn also showed a deficit for painters. These results were consistent with Mars = activity, Jupiter = outward show, Saturn = inner thought, which more or less fitted astrology. On the other hand there were notable conflicts that Gauquelin had listed side by side in L'Influence des Astres (p.212) and which can be summarised as follows:
Gauquelin's findings Astrological doctrineGauquelin comments: "The main reaction of those who examined the work was to advise me to try a replication of the experiments abroad, to see if the same results would appear among births other than French. I agreed. The scope of the experiments had to be extended if I wished to confirm the facts. Beginning in 1956, my wife and I visited a different country each year. During the holidays, and taking advantage of the thirty days of freedom at our disposal, we accumulated new records; during the rest of the year we worked on the statistical calculations. In this way German, Italian, Belgian, and Dutch celebrities were united. In all, 15,000 births were collected for analysis, all complete, indexed, and obtained from the registry office by means of an energetic correspondence with numerous City Halls" (Astrology and Science 1970:152-153). During this data gathering Francoise's language skills were crucial.
These 15,000 new cases, together with a further 2000 cases from France, led to the results published in his second book Les Hommes et les Astres. Denoel, Paris 1960, 268 pages. Writing it had taken Gauquelin six months of work without pay. But the results were almost uniformly positive. Of the total of 13 French planetary observations for eminent professionals (Mars, Jupiter, Saturn for physicians and scientists; Moon, Mars for champions; Moon, Mars, Jupiter for military men; Moon, Jupiter for politicians; Jupiter for actors; Mars, Saturn for Painters), no less than 11 replicated with statistical significance. Similar tests of non-eminent people in the same professions gave results close to chance expectation. It was an impressive achievement. Gauquelin's summary from p.161 is shown below:
Of 13 replications with foreign data All dataNote the marked difference between the chance and non-eminent results on the one hand, and the eminent results on the other. When the French and foreign data are combined (last column), 11 of the 13 results are significant at or below p=0.01, which compares with the 0.13 results expected by chance. To Gauquelin the results seemed conclusive: "Our research has given the same fundamental results across five different countries, for births during 1800-1930, in latitudes of 37 to 55 degrees spanning more than 20 degrees of longitude, despite frontiers, differences in language, institutions, mentality, or history. This is more persuasive than replications in the same country" (p.165).
In summary, the results using foreign data showed that the surpluses and deficits previously observed for Mars, Jupiter and Saturn at the birth of eminent professionals were real. The Moon (previously uncertain) was now linked to creative writers. Venus had no apparent effect, but was later found to reach significance when similar groups were combined, eg for writers plus journalists, and for actors plus musicians plus painters. As before, the results more or less fitted astrology, but the conflicts remained, especially the lack of an effect for ordinary people, which on Gauquelin's later figures meant 99.994% of the population. It was all very puzzling.
Nevertheless the results led to two general hypotheses, based on a total sample of 20,396 cases, which guided his work from then on: (1) Different professions are linked to different planets. (2) The relationship increases with eminence and disappears for non-eminent persons. The effect had nothing to do with signs or aspects. What mattered was the planet's diurnal position relative to the horizon -- whether it was rising or had culminated overhead. (The opposite positions below the horizon seemed also to be involved, but too weakly and erratically to justify routine inclusion, a view corroborated 30 years later by Ertel who found no consistent relation to occupation.)
Generally the surplus or deficit was between 10% and 25% of the expected key sector total. This may seem a lot, but about four-fifths of births fall outside key sectors, and in terms of the whole sample (which is what matters) the effect is much smaller and of no practical value, see later section on effect size. Detecting such small effects amid the sampling noise is difficult and is one reason why the Gauqueli findings were to become controversial. Nevertheless, for the Gauquelins, planetary effects had replicated and were now a reality.
Explaining planetary effects via heredity
"So once again I was on the hunt for dates of birth, but this time I was interested in the general population of men and women, and not members of an elite selected for a special purpose. In fact, planetary heredity had to be a general law of human nature, and no longer a perquisite of some people whose vocation was clearly marked. For five years I worked on the birth registers at several city halls in the region of Paris. In the total figure more than 15,000 matchings of parents and their children were collected [24948 individual birth data], which enabled me to calculate almost 300,000 positions of the planets" (Astrology and Science 1970:157).
The results repaid the huge effort and were published in L'Heredite Planetaire (Planete, Paris 1966, 226 pages), as if to herald the Gauquelins' own personal contribution to heredity, namely son Daniel born 22 April 1967. Parents with the Moon, Venus, Mars, Jupiter or Saturn in a key sector tended to pass the same planet on to their children, although not necessarily in the same key sector. Later Gauquelin explained how it works by reference to his own family:
"To take a practical example, my mother was born at Rouen in Normandy on 15 July 1900 at 4 o'clock in the afternoon, just at the rise of Jupiter. I was born on 13 November 1928 at 10:15 in the evening in Paris, at the culmination of Jupiter. In both cases, Jupiter was to be found in one of the key zones of the sky, and one could say that there is planetary heredity, as far as Jupiter is concerned, between my mother and myself. My sister was born in Paris on 9 October 1934 at 5 in the evening. Jupiter at that time was neither at the rise nor the culmination, and the hypothesis of Jovian heredity between my sister and my mother has not been confirmed" (The Truth about Astrology 1983:41).
The heredity effect was increased if both parents had the same planet in a key sector. There was no tendency to pass a different planet on, and there was no effect for zodiac signs or for aspects or for the other planets including the Sun. The strongest effect was for Venus, which supported the (weak) Venus effects recently discovered in eminent professionals. The heredity effect, like the planetary effect, was solely a diurnal effect and involved the same five planets. Although it was only half as strong it was still an important step forward.
In fact Gauquelin later considered planetary heredity to be the most important of his findings because it offered a clue for integrating planets with disposition and environment. It also showed that planetary effects existed for ordinary people as well as eminent ones. Heredity effects had the practical advantage of not requiring expectancies for their calculation, only the counts for the four combinations of parents (planet in key sectors yes/no) vs children (ditto yes/no).
How the planets worked was unclear. Gauquelin dismissed the idea that planets could somehow imprint something on the person, simply because the moment of birth is too late -- everything is decided at the moment of conception. So he speculated that the unborn child might have a particular planetary sensitivity that would trigger the birth process at a particular time rather than at some other time. The planet added nothing to the genetic makeup -- its position in the sky merely reflected what that makeup was.
To support this view Gauquelin pointed to the apparent disappearance of planetary effects for births after about 1950, which he attributed to the increasing incidence of interventions such as Caesarians that upset the delivery timing. On the other hand he noted that planets triggering birth would affect the beginning of labour rather than its end, but the beginning of labour for several thousand cases showed no planetary effect. As the first step to resolving this impasse, he wanted to study eminent professionals who had been born only via interventions. If the planetary effect disappeared it would strengthen his trigger idea. If it remained he would have to return to the imprint idea. He began such a study but died before it was finished (Neo-Astrology 1991:173).
Explaining planetary effects via personality
Fortunately for the analysis he had the help of Tom Shanks and the computer of Neil Michelsen's Astro Computing Service in San Diego. And by 1978 he had the results. They were positive. As he put it: "The relationship between planet and character trait could be observed without any need to take account of the professions. Persons with an iron will often had a tendency to be born under Mars, expansives under Jupiter, introspectives under Saturn and poetic temperaments under the Moon" (Truth about Astrology 1983:63). Furthermore the planetary effect sizes for character traits averaged 0.10, more than twice that for professions. It seemed to be an impressive vindication of Gauquelin's ideas. But there were snags.
First, a replication of the heredity effect in 1976 with 18,298 new parent-child comparisons was only weakly positive, with effects only half as strong as those in the first test in 1966. A third replication in 1984 with 33,120 parent-child comparisons showed no effect at all. In other words it was not clear whether the heredity effect (and thus the argument in favour of personality effects) actually existed.
Second, Gauquelin's trait calculations were based on the number of traits rather than the (much smaller) number of subjects. In other words his statistical results were based on inflated N's and were therefore much too optimistic. When repeated with correct N's their significance tended to disappear. Furthermore, the character traits were English translations of the original French, and had been extracted by Gauquelin after he knew where the planets were. So both procedures could have been affected by inadvertent bias.
Gauquelin accepted that a Gauquelin extraction bias had "very likely ... played a role in the [personality] studies", because computer checks of his manual results had confirmed the presence of Gauquelin bias in his sector calculations. But the calculation bias affected his results only slightly, so he felt it was "unlikely ... that [extraction] bias in the studies explains 100% of the results" (Correlation 10(2), 22, 1990). But biassing the translating of trait words (where there is always a range of English words for each French one) is easier than biassing the sector corresponding to a given birth time (where only a small proportion will be ambiguous). And in this case it needed only 1 in 12 translated traits to be biassed to produce the mean observed effect size of 0.10. So Gauquelin could be 11/12ths or 92% unbiassed, at first sight a level not worthy of suspicion, yet enough bias would remain to explain his results.
Third, Suitbert Ertel had made a study in which four specially-trained students extracted traits from the biographies of 110 eminent professionals. Ertel's approach avoided the above problems and his results showed no tendency for planets to be linked with the relevant trait (Correlation 10(2), 3-19, 1990). The extraction had also been done by Gauquelin with positive results, and by two students hired by Gauquelin, again with positive results. Gauquelin argued that Ertel's negative results were due to poor extraction.
However, this seemed unlikely because Ertel's extraction process had been carefully refined beforehand to minimise errors. Also, when Ertel compared the extractions for all three studies, it was clear that Gauquelin, not working blind, tended to extract more matching traits and fewer non-matching traits than did Ertel's students working blind, even though the students extracted on average twice as many traits as Gauquelin did from the same biographies. Furthermore, as noted by Ertel in Correlation 12(1), 2-9, 1993, Ertel's student extractions were used as is, whereas Gauquelin's student extractions were discarded by him if they showed poor agreement. Discarding was often a matter of interpretation, for example in deciding whether "active" was in agreement with "energetic", which of course allowed the same Gauquelin bias that the study was supposed to avoid. And once again there were more matching traits and fewer non-matching traits in the Gauquelin-selected student extractions than in Ertel's student extractions (the effect size of the difference was 0.23, p = 0.02, or quite enough to explain the supposedly positive results).
Ertel concluded "this is strong evidence for concluding that Gauquelin, as trait extractor, was greatly influenced by his knowing the sector positions of the persons whose biographies he worked on" (p.13), and that this was enough to explain Gauquelin's character trait results. In other words Ertel had shown that, contrary to what heredity required, the planets were linked with occupation, not trait. Gauquelin's objections could have been resolved by analysing his student extractions before selection, but he died before this could be done. For a detailed overview see Correlation 16(2), 10-39, 1997.
Psychology of the planets
First, for each of the ten planets, Francoise extracted their keywords from ten astrology textbooks selected for clarity and renown. The result was an average of 35 keywords per planet per textbook. With the help of Tom Shanks and Neil Michelsen's computer, she calculated each planet's sector distribution for professionals whose traits included a keyword for that planet, for example the Sun keyword power appeared in the trait lists of 225 professionals. She then calculated the significance level in key sectors for the combined keywords vs expectancy. The sample size was thus number of professionals (out of 1980) x number of traits (out of 35) and was typically about 21,000. She notes that the results are subject to semantic links between traits, so they are suggestive only.
According to Ertel's findings, the traits for each professional were inadvertently biassed by Gauquelin in favour of the planetary keywords for the Moon, Venus, Mars, Jupiter and Saturn, the result of Gauquelin knowing in advance planetary positions in key sectors and planetary keywords. So we expect Francoise's results for those planets to be positive -- and they are, although for Jupiter they are mixed, being positive for a bad Jupiter (extravagant, wasteful) and negative for a good Jupiter (generous, dignified). The mean effect size calculated from her results is 0.054, well in line with the Gauquelins' previous personality results and thus consistent with Gauquelin bias.
For the other five planets Gauquelin had previously found no planetary effects (and therefore no traits), so for them there could be no corresponding bias in the trait list, which makes Francoise's test a useful independent test. However, the results were negative, again in line with the Gauquelins' previous results. The mean effect size calculated from her results is 0.009, most of which is due to Pluto, probably a sampling fluctuation due to Pluto's sample size being less than one fifth of the others.
When Francoise examined the astrologers keywords for the five Gauquelin planets, she noted that a keyword "given by only one astrologer is most often contradicted [by the Gauquelin traits]", whereas one given by "more than one astrologer ... is most often confirmed" (p.57). Which is compatible with a Gauquelin bias arising from his reading many astrology books instead of just a few. On average 44% of keywords were on the Gauquelin trait list for that planet, 22% were on the trait list for a different planet, and the remaining 34% were on neither -- again a result compatible with the trait list having been shifted by Gauquelin bias in the direction of agreement with planetary keywords.
Setting an example
"For each case you must first find a reference source such as a biography or biographical dictionary. You must then copy the data without error, find the registry office address, and send off your request for the birth time on the birth certificate (which in France is available to anyone who asks for it) [and pays the fee]. When the data are complete you must check for summer time, convert to GMT, and calculate the positions of the planets, ascendant and midheaven. You are then ready to start tabulating, counting, and analysing, leaving only the worry of writing it up. The net result of all this is that the first one or two cases are a distinct pleasure. The next ten or twenty are not a pleasure at all. And by the first hundred you are going out of your mind. Only those who have done this kind of work will know what it takes. We owe more to Gauquelin than we know" (p.192).
Now take the business of explaining your procedures and results:
"First, there is a statement of the problems (of data, of astronomy, of demography, of statistical analysis) and their solution. Second, there are the results and conclusions. Third, and most important, there is concern about loopholes (does it replicate? are the controls adequate? what should be done next? what happens if ... ?) followed by work to provide the answers. This is how a good scientist proceeds -- by never quite believing what he is seeing until alternative explanations have been eliminated. A good scientist also explains everything clearly, gives full references, and anticipates our questions. And here Gauquelin never lets us down" (p.193). But it doesn't end there.
In their 1982 book Astrology Science or Superstition?, Eysenck and Nias put it this way:
"From the beginning he [Gauquelin] has had to provide his own funds for research, although he now does receive the occasional grant for his privately run laboratory. [Furthermore] his work has involved getting data from birth certificates in registry offices all over the continent, and for financial reasons he has had to rely on public transport for much of this work. Since registry offices are typically open to the public on only a few days a week, it can be seen that there were many difficulties to overcome. Also his research had to be fitted in with part-time work to earn money. In this respect he provides the complete answer to astrologers who point to the lack of funds to explain astrology's lack of scientific progress; clearly what astrologers need is not so much funds as initiative and determination!" (pp.183-184).
Of course over the years a few astrologers such as Krafft have indeed shown initiative and determination in carrying out heroic research studies. For example before 1950 in the USA the Church of Light laboured for 30 years and studied thousands of charts now in the ISAR collection, while the now-defunct American Astrological Association analysed over 100,000 birth dates, albeit both with methodology too unsound or too limited to make the results of value. Since then others such as John Addey, T Patrick Davis, Michael Munkasey, and Edith Wangemann have devoted 20 or more years of their lives to astrological research. Nevertheless the Gauquelin work is still the supreme example of rigour, initiative and determination that every astrologer can aim for but few will achieve. See Appendix 5 for a list of the main Gauquelin publications. For tributes to Michel and Francoise see elsewhere on this website.
Use of effect sizes
The use of effect sizes has many advantages. They simplify the otherwise complicated estimation of sample size requirements (the sample size required to reliably detect effect size r is roughly 10/r2). They are in common use, which allows comparison with effect sizes in other fields. And sets of effect sizes can be examined by factor analysis and meta-analysis. For example the use of effect sizes and (since the 1980s) meta-analysis in parapsychology is recognised as having sorted out a field noted for large numbers of conflicting results. And in astrology it is doing the same, see elsewhere on ths website.
Unfortunately Gauquelin never used effect sizes. Instead he generally presented his findings in terms of statistical significance, which varies roughly as N2 where N = sample size. As N increases, even the most trivial effect will eventually reach astonishing significance. For example suppose that a technique gives 51% hits vs 50% expected by guessing. With 100,000 cases it will give around 51,000 hits vs 50,000 expected by guessing, for which the significance level by chi-squared test is an amazing 0.000,000,0003, way beyond anything Gauquelin ever got. Yet the hit rate is still only 51%. In short, significance levels tell us little about how big an effect is. In Part 2 we will see how effect sizes lead to important insights that were missed by the Gauquelins.
Overview of 45 years' work in terms of effect size
Chart factor Professionals Parents vs ChildrenN is the mean number of professionals per sample or the mean number of parent-child comparisons. N is often very large.
An effect size of 0.04 is equivalent to tossing coins with a hit rate of 50(1+0.04)% = 52% instead of the 50% expected by chance.
For a client's chart reading, the minimum useful effect size is about 0.40, a hit rate of 70% instead of the 50% expected by chance.
* Occupation vs predicted sign. Traits vs predicted sign -002 (2206).
** The two Gauquelin replications gave .009 (N=18,298) in 1976 and -001 (N=33,120) in 1984.
The above table shows how effect sizes put Gauquelin's entire findings into context. For both professionals and families the mean effect sizes for signs and aspects are less than 0.005, and are often in the wrong direction, which shows that signs and aspects are without useful effect. That is, they do not mean what astrology books say they mean. In heredity, the number of children having the same sign or aspect as a parent is effectively no different from that expected by chance. For professionals the mean effect sizes for diurnal positions range from 0.035 to 0.047, mean 0.043 for 35 studies. They are replicable and highly significant, typically p=0.0001 for sample sizes around N=2000. But the hit rate of around 52% is of no practical use, especially as (on Gauquelin's figures) it applies only to the 0.006% of the population who are eminent. Applied pro rata to the general population, the hit rate is 50.0001% -- use astrology for a million clients and you might score one hit more than you would by tossing a coin.
Tests of astrologers
"It is useless to declare that astrology represents an infantile state of mind if well designed experiments prove that its pronouncements are correct. In that case it should once again assume an extraordinary importance. It matters little whether it is explained by symbolism or physics, or whether the stars are signs or causes. ... [But now] it is high time to come to an end. Every attempt, whether of astrologers or scientists, to produce evidence of the validity of astrological laws has been in vain. It is now quite certain that the signatures in the sky which presided over our births have no power whatever to decide our fates, to affect our hereditary characteristics, or to play any part however humble in the totality of effects, random and otherwise, which form the fabric of our lives and shape our impulses to action" (Astrology and Science 1970:125,138).
Astrologers predictably took little notice. Gauquelin comments:
"Astrologers remain unconvinced by the publication of these statistics totally destroying their doctrines. There was a time when they asserted, loudly and frequently, that statistics provided an ideal method of proving the truth of astrology, but this view was most popular during the period before the work of Choisnard and Krafft had been discredited. Today, astrologers have gone into reverse and claim, on the contrary, that statistics are irrelevant to astrology" (Astrology and Science 1970:136).
Gauquelin's response to the above U-turn was to test astrologers to see if they could match birth charts to people of opposite character such as long-lived vs died-in-infancy, or ordinary person vs four-times winner of the Tour de France cycle race. His matching tests pre-dated the 1961 matching tests of US psychologist and astrologer Vernon Clark after whom such tests are now named. But "astrologers regularly fail these tests and are sometimes so disillusioned that they accuse me of rigging the cases" (Truth about Astrology 1983:139). Since then such failures have been repeatedly confirmed by others including astrologers, so Gauquelin's failures had nothing to do with any supposed rigging.
Reactions of astrologers today
The second is to dismiss Gauquelin's results as irrelevant. For example James Holden says "his work was very narrow in scope" (History of Horoscopic Astrology, AFA 1996:236), as if Gauquelin had never tested what astrologers actually did or had no understanding of astrology. Not for nothing had Gauquelin read 100 astrology books while still a teenager! His Astrology and Science (1970) shows a stupendous scholarly grasp of astrology's history, principles, and related cosmic influences that has never been bettered in a single book. His Neo-Astrology (1991) contains what could be the most succinct description of astrological principles (and their disconfirmation) ever published. To dismiss his findings as irrelevant is absurd.
In fact an often-cited point made in 1979 by Arthur Mather is as relevant today as it was then: "Both those who are for and against astrology (in the broadest sense) as a serious field for study recognize the importance of Gauquelin's work. It is probably not putting it too strongly to say that everything hangs on it" (Zetetic Scholar 3/4, 96, 1979). Indeed, about 23% of all Correlation pages since it began in 1981 have been on the Gauquelin work, which confirms its importance in the eyes of researchers. (On the other hand, one might wonder why 23% of all Correlation pages should be devoted to a hit rate of 50.0001%.)
Baffling planetary puzzles
For astrology the puzzles include: Why only diurnal position and not signs or aspects, why traditionally weak positions (cadent houses) and not strong ones, why occupation and not character, why does the occupation effect disappear at low eminence and sometimes diminish at the highest eminence, why does it sometimes deviate from tradition (eg Moon instead of Mercury for writers), and why only five planets (Moon, Venus, Mars, Jupiter, Saturn) and not the Sun and other planets? Indeed, why an effect in the first place if isolated factors really are as meaningless as some astrologers claim? And why such a small effect?
For science the puzzles include: Why no link with physical variables such as size and distance (Figure 3), why no link with the sun, why is eminence important, why an effect only at birth, why contrary to all expectation is the effect larger for rounded birth times, and why does it disappear when the birth is induced or surgically assisted? Why is it enhanced if geomagnetic activity is high at birth or if both parents have the same planet emphasized? Since physical influences are present all the time, what is so special about those at birth, why is the link with birth and not with conception or the onset of labour? Indeed, why an effect in the first place? What evolutionary advantage could it confer?
For more than fifty years these puzzles have defied explanation. They seem totally and utterly baffling. Together with artifacts they are explored in Part 2.
Appendix 1. Michel on how it all started
We met in his [Gauquelin's] Beverly Hills hotel room. I found him to be a charming, thoughtful and critical-minded man who clearly enjoys his work. I was interested in learning how he became interested in astrology in the first place. "As far as I know", he said, "I have always been interested in astrology. I don't know why. My father was interested in astrology, but just for fun. He was a dentist. I learned to calculate a chart but when I was 16 I began to wonder, if maybe the whole subject as rubbish? So what could I do? I discovered some books by French astrologers who had tried to put astrology on a scientific and statistical basis. The first thing I did was try to replicate their work; I failed." (Fate June 1984:80-87, p.82).
Gauquelin gives a more detailed account entitled "Confession in the Form of a Prologue" in his Cosmic Influences on Human Behaviour (translated from the French), Stein and Day, New York 1973:17-20.As far back as I can go in my childhood, not to my earliest memories but to my first rational experiences, I have always known that I would be interested in astrology. Why? I do not know. Initially, chance played a role. No one in my family was an astrologer. They regarded horoscopes as a parlor game, nothing more. By the age of seven, I was asking my classmates their birth dates to tell them triumphantly the zodiac sign under which they were born.
By the age of ten, I was begging my father to show me how to calculate the ascendant of an astrological theme. This was during the "disaster" of 1940; my family, fleeing from the Germans, had taken refuge in the south of France for several months, leaving everything they owned behind them. However, I eagerly learned the laws of horoscope.
At the age of fifteen, together with a boy who has remained my best and most trusted friend, I cut classes at the Lycee Charlemagne to cross over the Seine and ferret through old Mr Chacornac's astrological bookstore opposite Notre Dame de Paris. We had no money to buy books, so we read them there in the store. By the age of seventeen, I had devoured more than a hundred works on astrology and I was drafting my own "Treatise" during the tedious hours of French-Latin classes. One day the teacher caught me at it. He read my youthful compendium and was quite disconcerted by it. Shrugging his shoulders, he returned my work to me and predicted that I would fail the baccalaureat examination, which was rapidly approaching. I have kept this notebook with its hard gray cover, my first "Treatise." Even this year, my friend and I were still laughing about his embarrassment. Yet basically, I had absorbed all the mysteries of the horoscope. A childish game, a waste of time -- who knows?
Until then, I had never asked myself this question: is all of this true? The limited success my predictions had earned me among my classmates, who had nicknamed me Nostradamus, and the small prestige which my specialty bestowed on me from very young persons of the opposite sex, were an ample reward. But the insatiable intellectual hunger that compelled me to read everything published about astrology was soon overcome by my desire to gather so many horoscopes that all skepticism would crumble under the weight of this proof.
I had discovered skepticism, not only in others, but also deep within myself. I could quote Descartes' first principle: Never accept anything as true unless I clearly and obviously know it to be true. But this would be deluding my readers and myself. One tends to rationalize afterward what has happened. I had at the very most a feeling that perhaps the astrological tree was concealing a forest of emptiness. Assurances from the astrologers I met were unrelated to the complex nature of the problem. It is true that for them there was no problem, and I found it increasingly difficult to tolerate their palaver. Was their proof only in their imaginations?
I decided to look for proofs and, if possible, to collect them. I began to frequent the Seine Archives, researching dates of birth. I filled notebook after notebook with figures, working as assiduously as a Benedictine monk. My head was full of such astrological aphorisms as: "Death occurs more frequently under the influence of Saturn;" "Professional soldiers are often born under the sign of Aries." I would find out if they were true or false. The little pocket money that I had was spent on stamps so that I could write to the many registry offices of France for records of hours of birth.
I was not the first to have this idea. Some astrologers, also claiming to be mathematicians, had believed that there was only one way to establish the validity of astrology: this was to provide statistics on a large number of births. [Here Gauquelin is referring to Choisnard, Kraftt and Lasson, whose claims were the focus of his first research.] Their books had convinced no one. Had they been mistaken? They had indeed been mistaken, for mere enumerations are not sufficient. One must know the laws of chance. I enrolled at the Sorbonne to learn them.
A feeling of hopelessness, at first vague, crept into my thoughts. It seemed that astrological laws were quite incompatible with modern scientific knowledge. This realization was reinforced by the very unencouraging results obtained from the columns of figures I had tabulated in my large notebooks. No, death does not occur more frequently under the influence of Saturn. No, professional soldiers are not born any more frequently than poets under the sign of Aries.
Yet the passion did not desert me for a single moment. I examined the most obscure aspects of horoscopes to drain the astrological abscess and burst the zodiacal bubble, but also perhaps with the secret hope of being that "hard-working hen" referred to by Kepler, which "in the stinking dung-heap of astrology may find a grain of corn, indeed a pearl or nugget of gold, if it searches and scratches long enough."
I scratched for a long time and eventually I found the gold nugget. At least I believed I had. But at the same time, I was very much aware how unlikely it was that this was true. Could my pearl be an artificial one, a slip of my thoughts, or a mirage conjured up by my subconscious? I was alone with my problem.
But fate was watching over me. On the university benches [this was in 1952] I met Francoise, my future wife. She was my first listener, my first reader and above all, my first critic. She advised me to write a book setting forth all the labor I had performed in secret. It was to be the "antiastrological" summary of my statistics and the nugget of gold. "And so," she said, "people will read you and criticize you. Then you will know if you have truly found something and if it is worth the trouble of continuing." At the age of twenty-five I wrote that book [L'Influence des Astres, finalised in August 1954, and published with Francoise's financial help in 1955] and I found that it was worth the effort to continue.
Thirty years later, in his The Truth about Astrology (Blackwell 1983:181), Gauquelin asks "Can there be a Conclusion?", and answers his own question as follows:Though I am so full of my subject [neo-astrology], so determined to defend it, so proud of my discoveries, I am still tormented by two feuding demons. The first is the fear of having been mistaken in asserting that astral influence is real; the second is the agonizing thought of all I have been unable to discover or explain. After 30 years of critical consideration of astrology, my passion for it has not diminished. But today I would not allow myself to draw drastic conclusions as I have sometimes done in the past. I will be content simply to have thrown a little light on this vast mystery which has occupied so many great minds over the centuries.
Nearly ten years later still, in his final book Neo-Astrology (Penguin Arkana 1991:178) published after his death, his concluding words are as follows:At the end of the twentieth century, two dreams must become reality: to go to the planets [space exploration], but also to know what the planets "do to us", so that as The Emerald Table says, "the miracle of unity may be perpetuated" [a reference to the unity expressed in "as above so below"].
Appendix 2. Francoise on how it all started
I was simultaneously baffled and filled with admiration. The topic of this research seemed senseless, but his objective methods seduced me. I longed for a husband who would accept my collaboration in scientific research and had found, until then, only wooers for a "stay-at-home," domestic wife. Michel accepted my participation in the research on his subjects. Without hesitation, I embarked on this adventure, certain that there was no risk involved. With objective methods, nothing positive could turn up from such a crazy enterprise. After a brief period of astrological verification, I hoped to convince Michel to abandon his projects for more quiet academic research within a normal curriculum at the Sorbonne. Our meetings developed into animated discussions about such things as how to properly divide the diurnal movement of the planets into thirty-six sectors, in which direction the sectors had to be numbered, and so on.
But my confidence in the quick effect of objective methods was too optimistic: the adventure proved to be much longer and much more difficult than I had expected. It has never ended.
Highlights from the adventure are recalled in her memorial to Michel in APP 8(1), 4-5, 1992 as follows:I first met Michel at Paris University where we were both students in psychology, at the time a still little known science in France. This allowed us later on to make a successful career as editors of series of books popularizing the main discoveries in psychology. But to Michel, getting to know better this new science was not an end in itself. It was rather a way to learn how to explore with correct scientific methods the astrological theories which had intrigued him and fascinated him since a very young age. [He had] tested the influence of zodiac signs, transits, aspects, with no results whatsoever. Also, innovative astrologers like Paul Choisnard, Karl Ernst Krafft, Leon Lasson had published books energetically asserting that they obtained significant results with their collections of data which Michel was unable to confirm with fresh data. This showed him how difficult it is to properly evaluate astrological hypotheses and induced him to look for proper training at the university.
But while following his courses in psychology and statistics, he began [writing] to French registrars for birth certificates of eminent professionals. To his amazement and delight, a first group of 576 Academicians of Medicine showed peaks of Mars and Saturn frequencies after rise and culmination ... It was only after much prodding by me that Michel revealed what kept him always so busy between courses and where his statistical explorations had lead him. Rather skeptical at heart, I offered him nevertheless my collaboration in this time and pocket-money consuming endeavor, provided we checked again with a new group of famous French doctors his first positive results with the Academicians of Medicine. At the Archives of Paris, we found a biographical dictionary listing the necessary dates and birth places of other French doctors, which, completed with each doctor's birth time, proved to be born with Mars or Saturn rising or culminating as often as the Academicians of Medicine had been. Very impressed by this replication, I became Michel's permanent coworker and soon also his wife. And we started checking all the professional celebrities available at the Paris Archives.
The results were plentiful: 570 French sports champions had Mars overfrequent after rise and culmination, 857 Academicians of Science had Saturn over-frequent in these areas, 676 Military Leaders had Mars and Jupiter over frequent in the same zones, and so on ... The five most visible bodies, Moon, Venus, Mars, Jupiter and Saturn turned out to be repeatedly significant in always the same way. These first French results were published by Michel in 1955. [Here Francoise forgets that the Moon and Venus were not added until later.] Then the same procedures were followed in other European countries during our summer vacations from 1955 to 1959. All the new replications of the French experiments provided identical results as the data collected in France. Later on, the same process was repeated also in the United States, again with identical results. ... While investigating other factors interfering with the astrological results (astronomical and demographic cycles) we had the opportunity to collect family data with which we tested Michel's heredity hypothesis.
Following Michel's death, Francoise continued publishing and editing Astro-Psychological Problems, which she had founded in December 1982 as a forum for the discussion of Gauquelin matters. Her many contacts within Europe, and her fluency in French, English, German and Italian allowed APP to have a notable European content even though all articles were in English. In 1995 ill-health forced her to discontinue publication, and eventually she moved to an old people's home in Joigny 130 km southeast of Paris, where she died in April 2007.
Appendix 3. Filling the experimental and methodological void
His preference for eminent professionals was soundly based. Earlier, in the 1920s, the psychologist EK Strong and others had spent many years studying various professions. They found that the professions were significantly different from each other in ways having no apparent connection with employment, for example in their likes and dislikes about people, hobbies, amusements, and books. The findings showed that a profession can represent a way of life as well as a way of earning a living, and led to what is now a widely used inventory, the Strong Vocational Interest Blank. Much subsequent research using this inventory showed that patterns of likes and dislikes do not arise from the profession but exist before a person enters it. Furthermore, they are as stable and permanent as any known aspect of personality. All of which confirms the soundness of Gauquelin's approach.
Having chosen to study eminent professionals, Gauquelin then laid down a method for collecting their data, starting with biographical dictionaries. For each dictionary entry without exception he extracted the date and place of birth, and wrote to the registry office at each place of birth to obtain the time and confirmation of the date and place. Initially he used only French births, and managed to collect nearly 6000, all supported by birth certificates bearing the town hall stamp and signature of the Registrar. All sources and the actual birth data are given in an appendix to his first book L'Influence des Astres (pp.245-343).
Now for the second void: rigorous methods of analysis. Once the data were in, Gauquelin calculated the birth chart, counted the frequencies of the relevant factors, compared the observed frequencies with the expected frequencies (corrected for astronomic and demographic effects), and then tested the differences statistically. A claim would be supported only if the difference was significant and replicable.Gauquelin began this work well before the advent of calculators, photocopiers, and home computers, so every letter, every record, every calculation, every tabulation, and every analysis had to be made by hand. The work could hardly be more laborious but it was nevertheless essential to the proper investigation of a controversial area. It had the huge advantage that the selection of data was objective and could be verified all the way from biographical dictionary to registry information. In this respect alone Gauquelin revolutionised the approach to astrological research.
Appendix 4. Early negative results
Heredity studies.  Aspects between the natal Sun of father and natal Sun of his children for 1873 French nobility (1113 sons, 760 daughters).  Aspects between the natal Sun of children for French nobility (first 500 comparisons).  Distribution of planetary positions in families for four generations (56 members) of the Aboville familly, and four generations (26 members) of the Hohenzollern family (the positions were compared with those for random control groups, but the latter were the more astonishing!).
Personality and aptitude.  Natal Sun positions (including signs) for 1995 generals, 1943 painters, 676 other military, 576 eminent doctors (Cancer was weak), 508 other doctors (Cancer was strongest), 500 actors, 623 criminals, 409 documented war deaths, 494 politicians, and 884 priests.  Natal positions of the other planets (Moon-Neptune), Ascendant and MC for 1154 eminent doctors, 884 priests, 676 military men, 500 actors, 570 sportsmen, 494 politicians, 1267 painters, 409 war deaths, and 623 criminals. At first sight the 623 criminals gave significant results for the Moon, with a correlation betwen split halves of r=0.45, p=.005, but among so many tests it could have arisen by chance.
Other studies.  Shunned aspects between natal Sun, Moon and natal others especially VE,MA,SA,UR for 623 criminals.  Aspects between natal Moon and transitting Jupiter (the astrological significator for good fortune) on the days 1154 eminent French doctors were elected to the Academy of Medicine.
Gauquelin comments that "This dry enumeration does not reveal the importance of the conclusions drawn from the results. They represent a considerable inquiry in the testing of astrological rules with large and varied samples. It is necessary to stress that the results demolish astrology more than they might appear ... because they attack not the claims of particular authors but the most elementary bases of the doctrine itself" (L'Influence des Astres 1955:62).
Appendix 5. Main Gauquelin publications
Methodes pour etudier la repartition des astres. LERRCP, Paris 1957
(English translation, Phenomena Publications, Toronto 1981).
Les Hommes et les Astres. Denoel, Paris 1960.
L'Astrologie devant la Science. Planete, Paris 1966.
L'heredite planetaire. Planete, Paris 1966.
The Cosmic Clocks. Regenery, Chicago 1967.
Songes et Mesonges de l'Astrologie. Hachette, Paris 1969.
The Scientific Basis of Astrology. Stein and Day, New York 1969.
Astrology and Science. Peter Davies, London 1970 (ex Planete 1966).
Cosmic Influences on Human Behaviour. Stein and Day, New York 1973.
Dreams and Illusions of Astrology. Prometheus, Buffalo 1979.
The Spheres of Destiny. Dent, London 1980.
The Truth about Astrology. Blackwell, London 1983. Same as:
Birth times: A Scientific Investigaton of the Secrets of Astrology. Hill & Wang, New York 1983.
Cosmic Influences on Human Behaviour. 3rd ed, Aurora, New York 1985.
Planetary Heredity. ACS, San Diego 1988.
Written in the Stars. Aquarian, Wellingborough 1988.
Neo-Astrology: A Copernican Revolution. Penguin, London 1991.
In 1984 Richard Nolle, book reviewer for Horoscope magazine, wrote "Every astrology student above the beginning level needs to know the whole story behind Gauquelin's research into natal planetary placements. [I judge] Birth Times [Truth about Astrology] to be the best single source on Gauquelin's work for the lay reader or astrology student." (Horoscope June 1984:41-42)
Gauquelin was also the author or editor of some thirty books on psychology.
Books and journals by Francoise Gauquelin include:Psychology of the Planets. ACS, San Diego, 1982, 110 pages.
Problemes de l'heure resolus pour le monde entier. Editions de La Grande Conjunction, Paris 1987, revised edition 1991, 360 pages.
Astro-Psychological Problems [APP] 1982-1995. Paris.
Of the thirty Gauquelin technical publications (totalling several shelf-feet of A4), seven are by Michel, one is by Francoise, and the rest are by both jointly. A comprehensive bibliography by Suitbert Ertel of the Gauquelin books and articles 1955-1991 appears in Correlation 11(1), 12-23, 1991. Perhaps the most concise overview of the Gauquelin work to 1992 is Ertel's "Update on the Mars Effect" Skeptical Inquirer 16, 150-160, 1992, includes 43 references and a survey of possible physical explanations. For more recent developments see Social Artifacts on this website under Gauquelin.