Competitions in astrology
Geoffrey Dean and Arthur Mather
Most of this article originally appeared as the following four articles by us in 1980-1987: £500 prize -- for you? (letter) Astrological Journal 1980, 22, 19-20. Did anyone win the world's biggest astrology prize? Astrological Journal 1981, 23, 162-166. Did anyone win the world's biggest astrology prize No 2? The results and a new bigger superprize Astrological Journal 1983, 25, 203-210. Superprize results Astrological Journal 1986, 28, 23-30, 92-96, 274-275; 1987, 29, 86-90, 143-147 (and also in FAA Journal 1985, 15(3-4), 19-32, and 1986, 16(1), 65-72). To improve on-screen readability, additional sub-headings have been inserted in bold italics, and repetitions (as when an article summarises an earlier article) have been deleted. Occasionally later material has been added in [ ] where this could be helpful.
Abstract -- The most common competitions in astrology are those inviting the reading of an anonymous birth chart, but they have little research value because they rarely have controls. More useful are competitions aimed at specific targets such as the Astrological Association's 1970-1980 £25 prizes for contributions to astrology, the £500 and £1000 prizes offered by Recent Advances in 1980-1981 for proof of signs, the $US5000 superprize 1984-1987 offered by an astrological consortium for evidence that the accuracy of chart interpretations cannot be explained by non-astrological factors, the 5000 Dutch guilders (about $3000) Astrotest of astrologers 1996, and the £200 Truth of Astrology competition 1997. Prizes for testing astrology have also been offered by US astronomers in The Gemini Syndrome 1979 (all fees and expenses would be paid for every success in mutually-agreed tests of astrological claims), and by the Centre Belge pour l'Etude Scientifique des Influences Astrales 1982 (a Grand Prix Astrologique of 100,000 Belgian francs, about $1700, for convincing evidence of a cause-and-effect relationship between heavens and terrestrial destiny). This long article describes in detail a total of nearly 70 entries (many representing a major research project) from a total of 15 prize competitions, of which perhaps the most notable and useful was the superprize. It was (and still is) the largest prize offered specifically in astrology, it was the only prize to focus on non-astrological factors, it attracted a good response from a total of 14 countries, and its panel of eight judges was easily the most expert panel ever assembled. The topics addressed by the entries were roughly equally divided between personality, events, and other areas such as discrimination, synastry and horary. Only one entry was successful but this was a fake entry designed to test allegations that the prize was unwinnable because appropriate tests could not be designed and the panel of eight judges was not impartial. This article brings together four articles by the authors originally published during 1980-1987. Although the competitions described are now more than twenty years old, they are unlikely to be repeated, and their results are as pertinent today as they were then. With 50 references.
The most common competitions in astrology are those inviting delineation of an anonymous birth chart. Examples from the 1980s and 1990s are T Patrick Davis's Astro sleuthing contests described on this website under Doing Scientific Research, Suzi Lilley-Harvey's "Who's Who" series in the Astrological Journal, and Charles Hannan's "Challenge" series in the NCGR Journal. Earlier examples include Alan Leo's tests of astrology 1906-1914 described on this website under Historical, and a 1927 competition in the USA described below.
Unfortunately such competitions have little research value because (1) there are no controls that allow non-astrological factors to be addressed, (2) even if entries were generated at random before ranking, something has to come top, and (3) entries usually show much disagreement, as in the 1927 competition. Here thousands of US astrologers attempted to win $US1000 (then roughly the average annual wage) by correctly describing three people from their birth data. But "they not only contradicted themselves, they were unanimously unsuccessful in describing the three people" (Miller R, Hugo Gernsback Skeptical Crusader, Skeptical Inquirer 26, 35-39, 2002).
Competitions relevant to research
When the prize is fabulous
In 1992 the Quebec Skeptics tested an astrologer of 25 years experience who claimed that her chart readings would be rated at least 75% accurate by their owners. Under an agreed procedure she prepared 3-page readings for each of seven anonymous subjects, who then rated all seven readings for accuracy on a scale of 0-100%. The mean accuracy rating of the authentic readings was 33%, more than the mean rating of 29% but less than the mean highest rating of 45% and well below the claimed 75%. Only one subject gave an accuracy rating of more than 75% but this was not for the authentic reading, which was rated at 28%. In response the astrologer said the poor result was probably due to the shortness of the 3-page readings, since her readings were normally of 20 pages. She added that she does not use clairvoyance or intuition, and that "if my work had not given positive results I would have left astrology long ago".
An earlier attempt on the Randi $10,000 prize was made in 1983 by US astrologer John McCall who, in an ad in the Washington Post in late 1982, took issue with claims that the Gauquelins had faked their data, announcing that "I also can demonstrate astrology in a scientific way". He challenged scientists to test his discovery and thus "prove once and for all that there is a scientific basis to astrology". McCall proposed the following test: "Let a person write down the date and place of his birth, plus four times of day. One of those times is to be his actual time of birth. By studying the face and build of that person and consulting astrological tables, I can pick out the true time of birth." His claimed hit rate was above 80% provided the births were not induced or Caesarean and the subjects were not non-Caucasian or senile. Randi provided under double-blind conditions a trial sample of five subjects that McCall agreed met all his requirements. But McCall scored only one hit, the result expected by chance (5/4), which he attributed to the small number of subjects. So a second test was arranged, this time with 28 subjects. McCall was completely satisfied with the test conditions, but he scored only 7 hits, again the result expected by chance (28/4).Gilbert D et al, Une fois de plus, l'astrologie echoue a un test [Once more astrology fails a test], Quebec Sceptique 23, 1-6, September 1992.
Randi J. A small-scale test of an astrological claim. Skeptical Inquirer 7(4), 6-8, Summer 1983.
Ianna PA and Tolbert CR. A retest of astrologer John McCall. Skeptical Inquirer 9(2), 167-170, Winter 1984-85
For insights into the baffling strangeness of some claims on these fabulous prizes see Randi J, Fakers and innocents: The one million dollar challenge and those who try for it, Skeptical Inquirer 29(4), 45-50, July-August 2005, available online as "Million Dollar Prize" at www.undeceivingourselves.com. Such challenges are not new, for example Joseph Dunninger (1892-1975), the famous mentalist and one-time chairman of the Scientific American Committee for the Investigation of Psychic Phenomena, offered $10,000 to any medium who could perform a feat he could not duplicate.
Another example occurred in 1870 when Christian flat-earther John Hampden offered a £500 wager (then seven times the average yearly income) to anyone who could prove the Earth was not flat. The wager was taken up by Alfred Russel Wallace, co-founder with Charles Darwin of the theory of evolution, with an experiment that involved looking along a straight ten-kilometre stretch of canal with a telescope. Points along the stretch dropped away as predicted by a curved surface, but Hampden insisted they dropped away in a straight line, so there was no curvature. In due course Wallace won, only to suffer a decade of threats and public abuse from Hampden that eventually put Hampden in jail. Ironically Wallace also lost standing among fellow scientists, who thought it unseemly to wager over an obvious fact because it made the obvious fact seem debatable. Today many scientists view the testing of astrology in the same light.Details of the wager are from Garwood C. Flat Earth: The History of an Infamous Idea, MacMillan 2007.
Targeting non-astrological factors?
The superprize has been briefly mentioned in several articles on this website, most substantially in Theories of Astrology under Applied Astrology. What follows is a detailed look at its entries and results, including those of the other prize competitions previously mentioned. The superprize began with the following letter from Dean and Mather in Astrological Journal 1980, 22, 39-40.
£500 Prize -- For You? [Letter]
The results are described by Dean and Mather in Astrological Journal 1981, 23, 162-166, as follows:
Did Anyone Win the World's Biggest Astrology Prize? [Article]
During our literature searches for Recent Advances and Recent Advances Two [still in preparation] we were therefore confident of finding such a demonstration. But although we had searched well over a thousand books and many hundreds of journals, and had written to hundreds of astrologers around the world, we had not found anybody who had demonstrated the validity of signs -- and many had tried. Accordingly, in the interest of filling the gap, we offered a £500/$1000 prize to anybody who could demonstrate the validity of signs. To ensure that the results were beyond doubt we specified that the demonstration must validate signs as traditionally conceived and applied (ie that Aries is assertive or whatever, and so on), must be capable of replication, must be supported by unambiguous evidence, and must be the entrant's own work.
This kind of reality for signs cannot be proven. The answer is why not? The astrological hypothesis is that each sign is of a different nature and that people born with a sign prominent in their chart will be of the corresponding nature. We can determine planetary positions in chants to high accuracy, and we can in principle measure personality and behaviour, hence a comparison can be made to see if the results support astrological claims. If the reality of signs cannot be proven, their effect cannot be detected. And if no effect can be detected, what is the point of using them? Or to put it another way, if they are non-falsifiable are they worth having? Which is not to say they cannot be legitimately used like a trait word to describe personality. What is not legitimate in the absence of proof is to claim that the hypothesis is true.
A sign is only one of numerous factors which in any particular chant can mask or even negate ifs effect. This kind of argument can be applied to virtually any variable in any discipline and if it were true we could never know anything about anything. The answer is either to use as test cases only those without such complications or to use a sample large enough for the complications to cancel out.
There was not enough time for evidence to be collected. The answer is that it is not unreasonable to suppose that the practising astrologer who uses signs will have already collected appropriate evidence (otherwise why do they use signs?). And among thousands of astrologers there should be some whose evidence already meets our requirements. Thus the time allowed is almost irrelevant.
Surely you were always sure that your money was safe? It would be safe only if a test of signs was not feasible or if we could cheat. But we have just shown that such a test is perfectly feasible [a point supported by the entries to our later invitation to devise tests of signs, see Response to an invitation on this website under Sun Signs], and we had an arbiter of fair play (see below). So safety is hardly a relevant feeling. In any case any truly rigorous validation of signs would be well worth the money.
How the Entries were Judged
Doreen Turnbull, Harrogate UK. A documentary on the harvest festival of a Mexican (?) forest tribe was seen on BBC TV about two years ago. The festival consisted of a series of rituals which Mrs. Turnbull linked with the signs of the zodiac. Thus the festival began with music and dancing (proclamation = Aries), masks were made (building = Taurus), masks were put on (communication = Gemini), and so on. Verdict: the entry fails because it does not demonstrate what is required, namely that people with signs prominent are of the corresponding nature. Furthermore, it does not meet the standards of proof specified by us, for example, the above sequence is not unambiguous because other attributions are just as convincing, such as free expression = Sagittarius, achievement = Capricorn, and secrecy = Scorpio, respectively. Mrs Turnbull subsequently added three more points to her entry: (1) The sign cycle applies not only to harvest festivals but to group activity in general. This point fails for the same reason as before. (2) Sign characteristics can be readily observed in individual character and behaviour. This point fails because no evidence was given that persons who are characteristic of sign X really do have the emphasis on sign X that astrology predicts. (3) She linked the zodiac to Teilhard de Chardin's idea of a psychic belt around the Earth which represents the collective thoughts and memories of mankind. This point fails because his idea is a hypothesis and not a tact. Also such a belt would necessarily be tied to the Earth (otherwise the Earth's motion would leave it behind) and hence would rotate with it, in which case it cannot be related to signs.
Rochelle Gordon, New York USA. Signs are valid but cannot be validated. Signs are symbols, i.e. they are words in the language of the universe. You cannot validate words, yet everybody agrees with their meanings and usage. Words work. Signs work. Verdict: the entry fails because if signs cannot be validated the prize cannot be won. The entry defeats itself.
Emma Drifon, Louisiana USA. I believe in signs. I read in Horoscope that the Sun conjunct Jupiter brings luck and money. At a horse race the No 5 horse won. Jupiter is No 5 sign (sic). No one picked horse No 5. Verdict: the entry fails because it does not meet the standards of proof required.
Dandon Campbell, Idaho USA. Changes in the Earth's angular momentum, as seen from the galactic centre, are divided by Mr Campbell into twelve categories which he then links with the signs of the zodiac. Verdict: the entry fails because it is a hypothesis, not a proof. Furthermore the Earth's angular momentum is constant, hence the changes on which the hypothesis is based do not exist. Mr Campbell subsequently related the signs to differences in travel direction between the Earth and Sun as seen from the galactic centre. This point fails for the same reason as before. Furthermore, when viewed from the galactic centre, the Earth and Sun always travel in the same direction, hence the differences on which the hypothesis is based do not exist.
Albert M Sterling, California USA. The entry consisted of two 45-minute tapes averaging one word every few seconds. Mr Sterling advised that it was against his rules to write (hence the tapes) but he has received the gift of knowledge and is fighting for the future of mankind. Astrology is a wonderful science and if we care to drop by he would provide some answers to our $1000 gimmick. Verdict: the entry fails because it does not meet the standards of proof required.
T Patrick Davis, Florida USA. This was easily the clearest, best-written and best-presented of all the entries and totalled 78 pages. Interpretations were given for 30 progressed charts, namely Queen Elizabeth II (seven events spanning 48 years), Winston Churchill (17 events spanning 90 years), and six others (one event each, usually death). The chart factors used were geo planets, helio planets (in the same geo chart), tropical signs, Koch houses, and eight major and eight minor progressed aspects of orb one degree. The interpretation was essentially traditional and was based mostly on the planet activated by the progressed aspect, the sign it was in and the corresponding house (ie the nth house for the nth sign), the sign ruled by it and the corresponding house, the house it was in, and the house on the cusp of the sign ruled by it. Traditional rulerships were used. Keywords were confined to a list of about 20 for each sign and house. In each case Mrs Davis showed that the chart matched the event and claimed that this was therefore proof of signs. Verdict: the entry fails because the interpretation was not shown to uniquely match the event in question. Thus in ber approach any given planet will on average be related to five or six different houses and hence to about half of the situations covered by the 12 houses. Because there is no situation not covered by houses, any given planet will on average be related to about half of all possible situations. Since each chart averages 18 progressed aspects, each fitting about half of all possible situations, the chance of not matching any given situation is (1/2)18 or less than 0.001%. Hence such matching proves nothing. Of course it may be that a standardised system of interpretation that permitted no arbitrary after-the-event selection would in fact demonstrate the validity of signs, but Mrs Davis did not use such a system. Because of the excellence of her presentation she wins the consolation prize of $25 for best entry.
A New Bigger Prize for 1981-82
1. The entry must demonstrate that the tropical sign hypothesis is true. The hypothesis in the words of Margaret Hone in her Modern Textbook of Astrology page 37 is as follows: "The astrological hypothesis is that each sign is of a different nature. People born with one or other of these signs prominent in their charts will be very much of the nature of these signs." In other words, the entry must demonstrate that each tropical sign is of a different nature, and that people with prominent signs, for example, those containing a preponderance of planets, are also above average in the corresponding nature. It is not enough to show, for example, that soldiers tend to be born under Sun sign X, or that lettuce grows best when planted under Moon sign Y.
2. The nature to be demonstrated may be taken from any standard textbook such as Hone, Mayo, Parker and Parker, or Tyl, or as summarised in Recent Advances page 122 [see below].Page 122 listed the keywords, needs, and traits (three positive, three negative) obtained from leading authorities (mainly Carter,
Hone, Mayo, Davison), and selected as being the most representative and best at distinguishing one sign from another, as follows:
AR=Leadership needs moderation: Bold energetic assertive, selfish insensitive aggressive
TA=Security needs enterprise: Practical conservative possessive, obstinate grasping fixed ways
GE=Relationships needs unity: Lively versatile restless, superficial erratic two-faced
CN=Emotional growth needs logic: Emotional protective sensitive, moody touchy unreasonable
LE=Self-expression needs humility: Proud magnanimous generous, domineering conceited shows off
VI=Perfection needs breadth: Analytical logical industrious, fussy interfering unemotional
LI=Harmony needs decisiveness: Pleasant harmonious tactful, indecisive untidy stubborn
SC=Power needs forgiveness: Intense passionate secretive, resentful vindictive obstinate
SG=Wisdom needs restraint: Active adventurous likes freedom, extravagant careless tactless
CP=Integrity needs sociability: Cautious practical persevering, selfish exacting narrow mind
AQ=Knowledge needs warmth: Detached original humanitarian, perverse eccentric low integrity
PI=Understanding needs drive: Sympathetic spiritual impressionable, confused impractical temperamental
3. The demonstration must be capable of replication, that is, others should be able to do the same thing and get the same result. Results which do not replicate do not demonstrate what they are supposed to demonstrate and hence do not qualify.
4. The entry must be supported by facts whose interpretation is unambiguous. For example, it is a fact that most people agree with descriptions of their Sun signs, but the interpretation of this fact is not unambiguous (most people also agree with descriptions of Sun signs not their own), and it certainly does not validate signs (it shows only that most people agree with descriptions that are truc of most people). Similarly a simple matching of charts to people is not unambiguous for the reasons given earlier. Entries accompanied by appropriate personality tests, statistical analyses and a replication are the ones most likely to succeed.
5. The entry must be the entrant's own work and must be submitted by the end of 1982 to Mr Zach Matthews BSc DFAstrolS (who continues as the independent third party to ensure fair play) at Oakfield, Goose Rye Road, Worplesdon, Surrey GU3 3RJ, England. However your intention to submit, plus a brief outline of your approach, must be sent to Mr Matthews by the end of 1981. [Zach Matthews, always a kind but critical supporter of our efforts, died in 2002.]
6. There is no race and all entries will be considered. If more than one entry qualifies then the prize of £1000/$2000 will be shared. If no entry qualifies then a consolation prize will be awarded to the best entry. Our cheque/check is ready and waiting. Good luck.
The results of this new prize (Prize No 2), together with a brief summary of the previous prize (Prize No 1), were given by Dean and Mather in Astrological Journal 1983, 25, 203-210, as follows:
Did anyone win the world's biggest astrology prize No 2? [Article]
Prize No 2
The rules required that the demonstration be capable of replication, be unambiguous, be the entrant's own work, and be submitted by the end of 1982. Notice of intention to enter was required by the end of 1981. As in the previous prize, Zach Matthews, editor of the Astrological Journal, was the independent third party to ensure fair play. For this report, he has kindly provided the following statement: "The description of the letters of intention received, and of the entries received and not received, is correct."
A few people felt that the words "prominent" and "nature" were vague and undefined. However, other textbooks were even more vague in defining what signs are, for example, The Only Way to Learn Astrology, page 10, says, "Each sign is a field of action in which the planetary forces operate". Other books, for example Basic Astrology: A Guide for Teachers and Students, and the Larousse Encyclopedia of Astrology, have no definition at all which could form the basis of a hypothesis. No doubt if we had attempted to impose definitions on Hone's wording we would have incurred the very accusation that we sought to avoid, namely of loading the dice. In any event, the entrants obviously did not feel this to be a problem.
T Patrick Davis of Florida, winner of the consolation prize for Prize No 1, gives in Mercury Hour (October 1981 pages 7-8) the reasons why she should have won. To her, the root problem was "the apparent lack of long-term experience in actual horoscope work by those judging the contest". Fortunately this will not be a problem to anybody else, since the judges included Charles Harvey, President of the Astrological Association and one of the UK's most experienced professional astrologers.
Her overall conclusion was that "the judges are not ready yet to conduct a truly meaningful and realistic contest to validate sign influences" and that "a prejudice in favour of a statistical approach ... will be a major hurdle to overcome". This misses the point. Her entry showed that charts match events, and in her eyes this should have been sufficient to win the prize. However, what she should have done, but didn't, was to check whether the matching could be explained by chance [which, as explained in our previous article, applies here]. Hence she did not win the prize. Of course, it may be that her approach could be shown to perform better than chance using methods other than the statistical approach which she eschews. In which case why didn't she do so? After all, those who feel that a particular scientific approach is inadequate are quite at liberty to devise something better.
Dr Mark Melton, a colleague of Pat Davis, argues in a two-part article in Mercury Hour (October 1981 pages 8-10, and January 1982 pages 49-51) that "the rules of these contests are slanted towards something which is not astrology", ie scientific demonstration, "and therefore the contests are rigged". In other words True Believers unite. Help stamp out Awkward Questions. He then invokes philosophical and terminological legerdemain to prove in effect that the prizes are impossible to win. A similar view is adopted by Ralph Holden of South Australia (Astrologers' Forum No 15 October 1982 page 4). However, there already exist studies which meet all the conditions laid down and which, if their results had been positive, would have won the prize (an example is the entry by Dwyer and Grange, see below). Melton also points out that the "astrological hypothesis is not a hypothesis at all but a generalisation of observable facts", yet "the astrological hypothesis is still not subject to proof by scientific method". In other words he is saying that observations cannot be observed. A curious view.
Results for Prize No 2
Of the remaining four intending entries, one by Ron Gaunt and Ira Smith of Queensland was not received and therefore fails. Their method involved selecting subjects with a preponderance of planets in particular signs, and then administering standard psychological tests. If the subjects exhibited the relevant traits significantly more than control subjects, if this was not due to self-attribution, and if the results could be replicated, this would be evidence for the reality of signs and the prize would have been won. That no entry was received suggests that such evidence was not obtained. The remaining three entries were received and were as follows:
Astrological Journal 25(2), 98-105, Spring 1983 and 25(3), 190-198, Summer 1983.
The number of authentic charts that were rated more accurate than the controls was 16 vs random and 21 vs 10-year; 22 random controls were rated more accurate than the 10 year controls. If signs have no effect, the expected number in each case is 17. By chi-squared test none of the results are significant at the p=0.1 level. Hence the subjects were unable to identify the authentic interpretations at better than chance level. If they had, if this was not due to self-attribution, and if the results could be replicated, this would be evidence for the reality of signs and the prize would have been won. But the authors rightly conclude that their results fail to validate the sign hypothesis. Verdict: the same, hence the entry fails.
Greig does not explain why image or persona should be shown by the Sun sign, when most astrologers relate it to the Ascendant, or why other planets are not important, but in view of the apparently positive results such objections are irrelevant. The real problem is that the matching was not performed blind. Numerous studies have shown that matching is strongly biassed by knowing the answer in advance. Hence we cannot tell whether Greig's results are due to astrology or due to bias (which may be quite unconscious). In other words, before we can accept Greig's conclusion we need to know whether the same results would be obtained under blind conditions. Since Greig doesn't tell us, we asked 20 astrologers to match his excerpts to Sun sign. The excerpts were in random order and excluded 13 excerpts which were too short to be meaningful, eg "masculine leading man". Responses were received from five astrologers, at least two of whom were full-time professionals.
According to Greig, up to 73 of the 108 excerpts should be correctly matched. In fact on average the astrologers matched only 14, range 11-17, thus supporting our suspicion that his results are due to bias rather than astrology. The average is still more than the nine expected by chance, but the difference can be explained by bias in the choice of star, bias in the choice of excerpt, and by direct recognition of the star from the excerpt (one astrologer correctly identified eight stars in this way). Of course, it may be that blind ratings of full biographies selected objectively would, as Greig claims, produce evidence of signs. But we can only judge Greig's entry as he presents it, not as he might have presented it. Verdict: the entry fails because the results are not unambiguous and therefore do not meet the conditions laid down.
The £20 consolation prize is divided two ways, half to Dwyer and Grange for the excellence of their experimental design, and half to John Greig for sheer hard work.
A new, even bigger superprize for 1983-84
These tests have attracted the same allegations as have the RA prizes, such as being rigged, irrelevant, and offering too little incentive for the effort required. Our response to all this is to devise a new astrology prize to accommodate these allegations. The details are as follows:
The superprize is not a race and all entries will be considered. If more than one entry qualifies the superprize will be shared. If no entry qualifies a consolation prize of 100 dollars will be awarded for best entry. If there are several best entries the consolation prize will be shared.
Aim of the superprize
There are no other rules. The interpretations can be of any kind such as character or life events, using any method whether recognised or not. Entries may be made by individuals or groups from any country. There are no restrictions on style, format or approach. In principle entrants can do whatever they like in whatever way they like.
Prof Hans Eysenck, University of London (psychology)
Dr David Nias, University of London (psychology)
Prof Ivan Kelly, University of Saskatchewan (educational psychol.)
Prof Marcello Truzzi, East Michigan University (sociology)
Michel Gauquelin, LERRCP Paris (cosmic influences)
Charles Harvey, President AA London (astrology)
Dr Henry Krips, University of Melbourne (philosophy of science)
[Dr Krips subsequently moved overseas and was unable to participate.
Prof Ulrich Mees, University of Oldenburg (psychology) kindly filled the vacancy]
A copy of each entry is sent to the three most relevant judges. If the entry passes two out of three, it will then be sent to another three judges for confirmation. The final outcome will be decided by consensus. The entrant will receive the judges' comments in full.
How to Enter
Guidance for Entrants
The outcome was described by Dean and Mather in Astrological Journal 1986, 28, 23-30, and in FAA Journal 1985, 15(3-4), 19-32, as follows:
Superprize Results Part 1
Full details and background information can be found in our superprize article published in astrological journals in England 1, Australia 2 and Germany 3, and in editorially abridged form in France 4, Holland 5, New Zealand 6, Canada 7 and the USA 8,9,10,11. Altogether news of the superprize probably reached 250,000 readers in the USA and over 5000 elsewhere.1. Dean G and Mather A (1983). Did anyone win the world's biggest astrology prize no 2?
The results and a new bigger superprize. Astrological Journal 25(3), 203-210, Summer 1983.
2. Ditto. FAA Journal 13(2), 32-38, June 1983.
3. Dean G, Mather A and Niehenke P (trans) (1983). 5000-Dollar-preis fur den Nachweis
der Gultigkeit astrologischer Deutungen. Meridian 3, 13-16.
4. Astro-Psychological Problems 1(4), 10-13, September 1983.
5. Wetenschap & Astrologie 6(1-2), 63, Winter 1982-83.
6. Journal of the Seasons 32, 17-20, Winter 1983.
7. Fraternity News 5(3), 33-34, Summer 1983.
8. American Astrology 2l & 53, July 1983.
9. NCGR Newsletter 4-6, Spring 1983.
10. Kosmos 12(3), 50-52, Summer 1983.
11. Aspects Summer 1983 under "What's Happening".
Over 60 intentions to enter were received from a total of 14 countries (USA 23, Germany 17, UK 9, rest of Europe 7, other 5). Of these 34 subsequently sent in entries. The combined entries totalled over 1500 pages or a pile 20 cm high. Organising the superprize required a total of three man-months of full-time work and $US350 for postage, stationery, printing and photocopying.
There were eight judges each chosen for their interest in astrology and expertise in related disciplines such as psychology. They were Charles Harvey, Michel Gauquelin, Dr David Nias, Profs Roger Culver, Hans Eysenck, Ivan Kelly, Ulrich Mees and Marcello Truzzi. This is easily the most expert panel ever assembled for the purpose of assessing astrological claims. Each entry was judged by the three most relevant judges. Thus a study involving signs would go to the judges most familiar with sign studies, and a study in German would go to judges who could read German.
Reason for the superprize
The answer is: a great deal. First, some of those interpretations turned out to be based on wrong charts, so the 100,00 perfect matches may only reflect astrology's ability to describe anything in retrospect. Second, there are many reasons why we should perceive validity in a reading even when none exists. For example the chart has so many indications it can't be wrong, we are so variable we would fit anything, the astrologer is so nice it has to be true, we project meaning where none exists like faces in clouds, we remember the hits and not the misses, the astrologer is recycling what we revealed 10 minutes ago, and we are both unaware that these things are happening. Third, to get around these problems there have to date been 7 blind trials of chart interpretations involving a total of nearly 200 subjects, and 20 blind trials of astrologers involving a total of over 400 astrologers and over 450 charts or pairs of charts, but not one has produced convincing evidence that authentic charts can be distinguished from controls [the numbers have since greatly increased but the outcome is the same, see Case for and against astrology on this website under Adroit Utilities, or Meta-analyses under Doing Scientific Research]. This is a lot of tests, and most of them were conducted not by hostile sceptics but by astrologers seeking to demonstrate the validity of their beliefs. Yet the results are precisely what we would expect if astrological factors do not work. In other words we can certainly examine books and visit astrologers, and thereby conclude that astrology works, but we cannot conclude that it works for the reasons claimed by astrologers, ie that it is factually valid. Which of course is potentially disastrous for the credibility and professional standing of astrology (is nobody the least bit worried?). The aim of the superprize was to change all this.
There but for the grace of God goes astrology
Phrenology provides a warning because it is so similar to astrology. Like astrology, it is concerned with individual potential and the philosophy of existence. That is, it encourages you to (1) think about yourself in phrenological terms, (2) assess yourself via phrenological principles, and (3) act on the findings to achieve a physical, mental and spiritual whole and thus harmony with the world. So astrology and phrenology cover the same ground. Like astrology, phrenology attracted people of intelligence and standing, and achieved even greater popularity, for example a reading was often a prerequisite for getting a job. As in astrology, theories were expressed in tones of complete authority despite a complete absence of experimental evidence (the workings of the brain were then largely unknown, as were experimental techniques for testing phrenological claims). As in astrology, learned arguments raged in the Phrenological Journal on topics that ought to have precluded argument, for example some held that head size indicated intelligence while others disagreed, and every criticism of phrenological ideas was furiously attacked. As in astrology, famous people were quoted in its support. For example, phrenology is "as much better than all other systems as the electric light is better than the tallow dip" (Gladstone), and it has "an accuracy that the most intimate friends cannot approach" (Alfred Wallace FRS). But the most important point is this: like astrology, phrenology persisted because day after day its practitioners and clients could see that it worked. It was "so plainly demonstrated that the non-acceptance of phrenology is next to impossible". So no matter what the critics said, they knew. Altogether a perfect parallel to astrology.Quotes are from Severn JM. Popular Phrenology, Rider, London 1913, p 6.
But the fundamental postulates of phrenology were wrong. Dispositions are not indicated by size and shape because the brain does not work like that. Nor do they exist as the fundamental entities required by phrenological theory. So a bulge here or a depression there cannot mean what it is supposed to mean, even though the underlying philosophy of "know thyself' has undeniable appeal. As a result, phrenology remains "the most popular of all the doctrines .... (in the history of psychology), and at the same time the most erroneous. It affords a striking example of the danger of erecting a vast superstructure on inadequate observation and inexact methods".Quote is from Flugel JC. A Hundred Years of Psychology, 3rd edition, Duckworth, London 1964, pp 3l-38. A wonderfully sympathetic but devastating critique of phrenology that applies equally to astrology. All astrologers should read it. More details of the evidence against phrenology are given by EG Boring, A History of Experimental Psychology, 2nd edition, Appleton Century Crofts, New York 1950, pp 50-79. [Subsequently a thorough review of phrenology vs astrology became available, see Dean G. Meaningful coincidences: Parallels between phrenology and astrology Correlation 1998, 17(1), 9-40.]
Clearly, phrenology did not need to be true in order to work. To put it another way, phrenology was like an uplifting discussion in which only white lies were exchanged. Critics saw only the lies while practitioners only saw the uplift. Both were right -- but since lies are unethical, phrenology inevitably failed to survive. Astrologers will of course recognise this as a fitting end to a system whose breakdown of human experience (into commonly 42 divisions spread across 8 categories) is at such variance with their own 12-fold system.
But astrology is in the same boat. As in phrenology, it is evident that astrology does not need to be true in order to work -- if invalid factors worked for phrenology then they will work for astrology. The same confusion exists between lies and uplift. And the arguments currently made in support of astrology apply equally to phrenology. Which brings us to the bottom line. Astrologers are like phrenologists: their systems cover the same ground, they apply them to the same kinds of people, they turn the same blind eye to the same lack of experimental evidence, and they are convinced for precisely the same reasons that everything works. But the phrenologists were wrong. So why shouldn't critics conclude for precisely the same reasons that astrologers are wrong? Read this section again, think about it, and let us know! [There was no response.]
What entrants had to do
If an entry was clearly deficient in design or reporting, Dean advised the entrants of each deficiency (even though under the rules there was no obligation to do so) and suggested that to improve their chances they might like to amend the entry before submitting it to the judges. About half the entrants were grateful and acted accordingly (some subsequently won the prize), the rest angrily denied there was anything that could be improved (none subsequently won a prize). Overall, as we shall see, the entries provide a fascinating array of research ideas and pitfalls for the unwary, and give us perhaps our best view yet of how astrologers tackle astrological research.
Comments and criticisms
On the other hand we have many critics. For example Dennis Elwell 12 argues that the prize is unwinnable because it is difficult to design an appropriate test, and because some judges may have too much at stake to be impartial. Barbara Koval 13 argues that science is not able to determine the validity of astrology because "Science deals with averages and groups. Astrology deals with individual experience." Jeff Jawer 14 argues that the superprize offers a "stacked deck" because he doubts the ability of some judges to be impartial. Mark Melton 15 argues that experiments cannot be designed that would win the superprize, hence it is rigged and unwinnable. Al Morrison 16 took a full-page ad in Mercury Hour to ridicule the superprize and to claim that it had been boycotted. In a libellous update 17 he claims that the superprize is moot because the sponsors "guaranteeing payment of the prize ... could not pay out such sums".12. Superprize debate between Dennis Elwell and Geoffrey Dean, Astrological Association Conference, York 1983.
Available from the Astrological Association as Sagicassette Y14.
13. Koval B. Letter to the Editor. American Astrology pp l8-19 September 1983. See subsequent issue for Dean's reply
14. Jawer J. Comments of the superprize. Fraternity News 5(4), 51, Fall 1983.
Dean's reply appeared in 6(3), 39-42, Summer 1984
15. Melton M. Mercury Hour l7-18, April 1984 and l-2, Extra Edition April 1985.
Dean's replies appeared in Mercury Hour 4l-42 October 1984 and 1985 in press [October]
16. Morrison AH. The alternative Geoffrey Dean PhD space age bonanza super-duper prize(s). Mercury Hour 19, January 1984.
17. Ditto. CAO Times 5(4), 43-44, 1984
These criticisms boil down to two points: (1) Appropriate tests cannot be designed. (2) The judges are not impartial. The first point is completely disproved by the entries.
The second point was put to the test by sending a 17-page control entry to all judges. The control entry was disguised as a genuine entry and was purportedly a study of transits that showed a strong and highly significant positive correlation (r=0.5) with life events. To avoid clues about its real origin it was prepared and mailed with the help of Astrosearch in Sydney NSW; copies are available on request. [See appendix for an abstract of this fake entry. No requests for copies were received.] It included all the things that a good entry should have, such as a literature review, adequate sample size, controls, statistical analysis, and a replication. It was in fact based on a genuine study, the only difference being a gross inflation of the sample size and the substitution of strikingly positive results for genuine negative ones. It was not entirely flawless (too good a result might have aroused suspicion) but it was certainly good enough to merit approval from an impartial judge. According to the critics our judges should have found some excuse to condemn it. This is what the judges actually said:Culver: Satisfies the requirements for clarity and chart interpretation.
Eysenck: Cannot fault the study presented. The results would certainly qualify for the superprize.
Gauquelin: Interesting entry. The best I have read. Deserves encouragement.
Harvey: (No comments despite a 6-month wait.)
Kelly: A promising study...the authors are aware of confounding variables that are a problem in this area.
Mees: Very impressive...the only work that...has been fully convincing...I hope it succeeds in the astrology superprize.
Nias: The best study so far, with replication and documentation of life events as bonuses.
Truzzi: An impressive...quality study... demanding independent replication.
You can see that the control entry easily achieved the minimum of two assenting judges required to qualify for the superprize, so in principle it could have won. Some judges provided a page or more of constructive comments. One even offered to help with a replication. Such actions and the above comments should reassure any critic inclined to doubt the judges' impartiality. Ironically the only judge to remain silent was the astrologer.
Entries not submitted (25 cases)
Bevan* Norway. A test of Kundig's rectification method using subjects with accurate birth times with or without a Void-of-Course Moon. If the method works then so does astrology. In which case if the method fails for subjects with a Void-of-Course Moon, then Void-of-Course Moon and signs work. Postponed due to pressure of other work. Will be finished and published in due course.
Dwyer* UK. Can subjects pick authentic interpretations from controls? A test of Starword computerised chart interpretations based on signs and aspects; these factors were used because tests had shown them to be the most reliable. See Dwyer T. A new approach to natal interpretation. Astrological Journal 25(2), 98-105, Spring 1983 and 25(3), 190-198, Summer 1983.
The control interpretation was calculated for a birth time 5 years and 6 months before the authentic time, except that the sun sign was kept the same to avoid giving clues to subjects familiar with their sun sign meaning. Starword synthesised all chart factors via their traditional meanings and took 45 minutes to produce each pair of interpretations. Each interpretation was divided into 7 sections (eg mentality, career, relationships etc) which the subject rated for accuracy. Work was abandoned after a pilot study of 30 subjects showed that 15 picked the authentic interpretation as the more accurate, and 15 picked the control, which is the result expected if astrological factors do not work. This is an example of a simple test that meets all requirements. Thus all non-astrological factors are automatically controlled in one hit; and the possibility of positive results being due to ESP rather than astrology is easily controlled by leaving the interpretations blank. Yet according to many critics such tests are impossible to devise. If this pilot study had produced the positive results expected, a larger follow-up study would have been conducted. If this had also produced positive results, the superprize would have been won.
Forneret USA. A test of charts vs life events.
Freemantle Swaziland. Can astrologers identify homosexuality? A Vernon Clark test with at least 10 pairs of charts and at least two astrologers. Test was abandoned due to lack of interest from homosexuals and astrologers.
Groeger* Germany. A statistical test of rectified charts vs life events for 200 subjects.
Hawley* UK. A statistical test of progressed composite charts of Cruft champion dogs and their owners at time of first success.
Heop* UK. A test of 50 students who are changing their life direction via a residential course. Do they have more contacts (eg Uranus transits) than a control group no changing direction?
Hoen and Hepp* Netherlands. A demonstration that subjects sharing a common condition, eg occupation or illness, have the same astrological signature in their charts. The signature consists of a multiple-of-22.5° aspects between the AS or M C and relevant planets, their midpoints or antiscions. Unfortunately the work could not be completed in time. The occupation chosen was astrology. The subjects were 100 astrologers whose birth data had been published in Zenit in the 1930s. The main finding was that 99 of the astrologers had a close hard aspect between AS, or MC, or their antiscion, and the midpoint of Uranus and the 4th/10th ruler, whereas it was present in only 30-35% of a control group of unspecified size. This, of course, may merely be an artifact of the large number of factors tested. Results will be published when ready.
Kern Germany. A demonstration that his chart works for character and events.
Kroncke* Germany. A demonstration that subjects sharing a common condition, eg occupation, personality or accidents, have the same astrological signature in their charts. The signature consists of specific Uranian planetary pictures.
Kuhnert Germany. Not clear but seemingly a test of charts vs events.
Landeau USA. Are differences between the geo and helio positions of Mercury, Venus and Mars related to personality?
Lavastre France. A demonstration that her chart works for character, health and events.
Methorst et al* Netherlands. No details but possibly involving signs.
Naiman* USA. A demonstration that charts indicate illness. The basic idea is that illness is caused by stress, and stress is visible in the chart. The work was already complete when the superprize was announced and is available as a 7-volume monograph on medical astrology. It was not submitted due to a mistake in the rules made by the editor of the journal concerned.
Plooyer* USA. A statistical analysis of aspects between married people.
Pritchett* UK. Can astrology identify personality? A Vernon Clark matching test using 5 astrologers and 30 pairs of subjects greatly differing in personality. Interrupted by the addition of twins to the family. It is hoped to continue after a year or so. [No report has been subsequently published.]
Rieber-Mohn Norway. Not clear but involving the analysis of charts classified by sub sign and by time period.
Rivers USA. A one-case study showing that astrology can distinguish between murder and suicide.
Rudolph Germany. A demonstration that astrology can solve questions of disputed paternity if the birth data of the possible fathers is known.
Schure* USA. A demonstration that charts indicate events for 10-30 subjects each with certified birth times. Abandoned due to work problems caused by the US recession.
Sprayer Italy. Solar chart interpretations for subjects undergoing knee surgery.
Staley et al* Canada. (1) An analysis of chart factors at the time of psychiatric hospitalisation for patients and controls. (2) The correlation between chart factors and 16PF scores for a large sample of nurses.
Swain et al* USA. Can astrologers predict marriage and the birth of children? Vollmann Germany. Not clear, no details.
Entries submitted but not judged (16 cases)
Christina USA. One page claiming a correlation between Moon phase and hair growth after a haircut. To stimulate growth, cut at New Moon or when Moon is increasing during First Quarter.
Cohen USA. 17 pages on numbers and their significance in astrology. Proposes that the degree and minute of chart placements indicate the age at significant events. For example her MC is at 11 Cancer 28.9, when she was aged 11 years and 5 months her sister was born, and at age 28.9 her sister came to live with her.
Darlak USA. 10 pages on why astrology is important in 1984. Mostly assertions and anecdotes plus homely advice.
Dunham NZ. 10 pages describing six case studies that confirmed astrological predictions.
Dwyer NZ. 154 pages on how to survive the next Aquarian Age. Mostly transcribed lecture notes based on Cayce and the Kabala and frequently impossible to follow. Example: "Pluto in traditional astrology up until a few years ago seemed to be ignored except to say it was generation influence but them are not all individuals part of generation was accepted they said that Pluto is the force of re-generation"(sic). Included 9-1/2 hours of cassette tapes. Contained nothing obviously addressed to the required topic.
Eaton USA. Eleven pages summarising the factors and techniques found to be reliable in 14 years of experience. Included biorythyms, palmistry and acrophonology [using symbols to represent sounds].
Julian USA. Eight pages of chart interpretation for a client plus 11 pages on astrological symbolism
Milsten USA. Five pages of speculation that astrological charts can be interpreted in electronic terms. Thus Sun generator, Moon battery, Mercury frequency tuner, Venus AC, Mars DC, Jupiter amplifier, Saturn resistor, Uranus flow of electricity, Neptune static electricity, Pluto magnitude, and Earth magnet. Each planetary influence increases progressively from Aries through Pisces on a scale of 1 through 12. A typical brief interpretation is: "Small reserves of energy but can generate needed energy to meet challenges. More inclined to flight (AC) than fight (DC). Fights by inertia rather than action."
Nielson USA. Three pages of interpretation for two charts. The aim of the interpretations was to convince the judges of the validity of astrology.
Oster Germany. 264 pages consisting of (1) Chart interpretations for 17 great composers, seven great physicists, Napoleon, a murderer and his 22 victims. (2) Essays on heredity, the law of existence, and the conception chart. (3) A report on blind judgements made by the author on two charts in 1946; an independent panel found him to be correct, however, the charts and details have been lost.
Patching Australia. 15 pages and 24 charts on why psychic astrology works, plus 60 mimeographed pages on the signs and zodiacal ages. The author was converted to the psychic zodiac of 13 signs after reading James Vogh's Arachne Rising, and has reached the conclusion that traditional astrology "as practised today is irrelevant or wrong". Her ideas came in a burst of psychic revelation.
Powell USA. Five pages and one horary chart for the question "where is my lost jewellery". Author implies that the jewellery was found but does not specifically say so.
Richards UK. 32 pages describing appropriate astrological indications for 40 events in his life. Beautifully drawn and presented.
Scholtz Germany. 85 pages describing 22 case studies (mostly events) using 10 planets, Transpluto, Koch houses and midpoints. Often esoteric. Beautifully drawn charts.
Seger USA. Five pages describing two chart predictions of accidents that came true. Both predictions involved progressed contacts using orbs of 2 degrees. No controls, statistical checks or awareness of non-astrological factors. In a subsequent letter the entrant says that if the judges "fail my entry for any reason, there is either something wrong with their ability to recognise bona fide astrologically-caused events, or with their own interpretational methods". Unfortunately he inadvertently misfiled his letter and retrieved it too late (mid 1985) for this point to be put to the test.
Titsworth USA. 19 pages describing 12 case studies that confirmed Church of Light rules for events.
Entries withdrawn or with negative results (5 cases)
Hegeler Germany. Not clear but involved the I Ching. After one year "Mein Plan war fehlerhaft", her plan came to nothing.
Lopez Mexico. 167 pages. Hypothesis: MMPI scores indicate the presence of certain factors in a chart. For 191 normal subject aged 17-55 in Mexico City a multiple discriminant analysis was performed between each of the 13 MMPI scores and the presence or absence of each of 100 chart factors. The factors were sign-related positions such as sign, element and decanate for 10 planets and Ascendant, and multipe-of-300 aspects (orb 15 degrees) for all planetary pairs except Neptune-Pluto. Of 13x100 = 1300 tests, 54 were significant at the 0.05 level vs 0.05x1300 = 65 expected by chance. The difference is not significant (p=0.3 by 2x2 chi-squared test) and is in the wrong direction.
Lutz Germany. Five pages. Hypothesis: given the dates of 6-10 important life events, the author's semi-Kundig method can determine the birth time within 1/2 hour. For 10 examples supplied by us, the author's calculated time differed from the known time by an average of 9 hours, the distribution of differences being almost exactly at chance level. Only one result was within one hour, which is no better than the result expected by guesswork.
Swick USA. Seven pages. Hypothesis: there are significantly more quincunxes at death for authentic charts than for control charts. Quincunxes were counted between nine transitting factors and 12 natal factors, orb 4 degrees, for 100 accidental deaths obtained from Las Vegas newspapers 1983-84 and 100 control charts on three dates chosen at random. Expected mean = 4.80. Observed means = 4.74 for the deaths and 4.57, 5.52 and 4.62 for the controls. The hypothesis was therefore not confirmed.
To assist judgement each judge received the opinions of all the judges, and no consensus was required -- the recommendation of a single judge was sufficient to award a consolation prize. In fact four of the six winners were chosen by more than one judge. The winners of the $US200 consolation prizes are Ferguson, Klein, Niehenke & Boer, Schneider-Gauquelin, Sohnius, and Urban-Lurain, all of whom have been sent their prize money. This leaves a balance of $US2300, plus further pledges from ISAR, Marguerite dar Boggia and Matrix totalling over $US400, which the sponsors have agreed to apply to a new prize as described below. Thus although the sponsors could have simply declared a stalemate and paid out the minimum (2%) provided under the original rules, they have actually paid out 35% and have committed the rest to furthering astrology. This should reassure critics like Al Morrison who claimed the sponsors were not bona fide. All entrants have been sent the judges' original comments on their entries.
The Astrological Association have preferred to stay with the original rules. Accordingly the superprize will continue under the original rules, using their portion of the prize money, and the results will be announced in due course. At the time of writing a winner according to the original rules is still possible. One of us (Mather) would have preferred more sponsors to have stayed with the original rules, but under the circumstances recognises their right to resolve the immediate dilemma while providing a constructive approach to future prizes. [There was no response to the AA's continuation of their portion of the superprize.]
Other astrology prizes
The Grand Prix Astrologique of BF100,000 (about $US1700), announced in the Belgian astrological journal Demain in 1982, was to be awarded for convincing evidence of a cause-and-effect relationship between heavens and terrestrial destiny -- a topic that makes the superprize look like child's play. Not surprisingly the response was too poor to justify awarding the prize. However BF25,000 has been made available to assist future research by one entrant (Nick Kollerstrom) provided that it follows the judges' recommendations. It is hoped to re-run the prize in due course 19. [No re-run has been announced.]
In 1970 the Astrological Association announced that it will award annually a prize of £25 (then about $US60) for the most valuable contribution to the study of astrology. A second prize of £10 may also be awarded in special circumstances 20.18. Culver RB. Personal communication August 1985.
19. Brahy GL. Personal communication August 1985. See also Correlation 3(2), 2, November 1983 and 5)1), 2, May 1985.
20. For rules see The Astrology Prize, Astrological Journal 12(l), l2-13, Winter 1969-70.
Details of each entry for the 1970 prize are given in The Astrology prize 1970, Astrological Journal 12(4), 5-7, Autumn 1970.
Details of later prizes usually appear in Autumn or Winter issues, or very infrequently in Transit.
All entries had to be typewritten, anybody could enter, and the entrance fee was £1. The 1970 prize attracted nine entries, but the three judges (astrologers Davison, Firebrace and Mayo) reported that their general standard was a little disappointing in that very few made a positive contribution to the study of astrological -- only two had any real claim to originality. Unfortunately, in neither of these cases was the presentation really persuasive, nor were practical examples appended in support of the theories advanced.
The winner was Harold Wigglesworth for the astrology of towns and cities, and a special award was made to the entrant who submitted a labour-saving MC-and-Ascendant finder. In 1971 there were 6 proposed entries but only two were submitted. According to the judges (astrologers Lind, Matthews and Russell), the first entry on planetary numbers lacked relevance to astrological study, and the second entry on historical cycles lacked background and supporting evidence. Neither won a prize. In 1972 there were two entries, both of which were later withdrawn. In 1973 the prize was won by Stephen Arroyo for his MA Thesis on Astrology vs Psychology [subsequently published as part of Astrology, Psychology, and the Four Elements, CRCS 1975]. The announcement occupies only four lines and gives no further details or even the number of entries, so presumably the response was poor.
Nothing further happened until 1979, when there were three entries. The first prize was won by Michael Mayer for a holistic perspective on meaning and identity [later expanded to The Mystery of Personal Identity, ACS 1984, see Using astrology as a counselling tool on this website under Applied Astrology], and the second prize was won by Michael Startup for an empirical study of Jupiter vs Saturn in the diurnal circle (the results were at chance level). In 1980 the prizes and entrance fee were doubled, but the response was so poor that no first prize was awarded and plans for future prizes were cancelled. For comparison the 1980 ($US1000) and 1981 ($US2000) RA prizes for the validation of the signs attracted 6 and 4 entries respectively, but no winners. The response to the superprize has therefore exceeded the response to all the above prizes combined.
Interestingly the response to astrology prizes today seems to be on a par with the response to phrenology prizes nearly 150 years ago. Thus in 1838, when interest in phrenology was at its peak and £1 was worth the equivalent of £10 today , three 5-guinea prizes [£5 5s] offered in Glasgow for the best essay on phrenology attracted a total of 11 entries, all "very able productions". And in 1839 a £20 prize offered by the Phrenological Journal for the best essay on choosing members of parliament by phrenology attracted 9 entries totalling 200 pages, some of them having "passages of considerable merit".From Phrenological Journal 1838, 11, 90 and 1839, 12, 415. No other UK prizes are mentioned during the period 1828-1847
[Here the test of astrologer John McCall has been omitted because it was described earlier.]
A new prize to challenge the critics
Nevertheless at the 1983 superprize debate between Elwell and Dean 12, critics claimed that the superprize had got it all wrong and that it ought to have been done differently. (Interestingly the worst hostility came from those opposed to any idea of testing their astrological beliefs, which seemed to raise doubts about their confidence in those beliefs). Accordingly the sponsors have agreed to make the remaining money available for a prize to be devised and administered by the interested critics, so the latter now have a chance to put their ideas into action. In other words the sponsors provide over $US2700 in prize money, the critics do the rest. What could be fairer? The only provisos are that the new prize be addressed to validating everyday practice, that it has the approval of the sponsors, that entry details be published in this journal before the end of 1986 and that the results be published in the same detail as here. Otherwise the offer lapses. Of course some critics may feel that the provisos are too restrictive and that a completely fresh approach is justified, for example perhaps the prize should be offered for any proof of astrology, or perhaps there should be more sponsors to increase the incentive, or perhaps smaller prizes should be offered for proposals for a definitive test of astrology, in which case their ideas are most welcome.
Interested critics should immediately circulate there proposals to the sponsors direct. Addresses may be obtained from: Geoffrey Dean, Box 466, Subiaco, 6008, Western Australia. We await the results with interest. [There was no response, so the offer lapsed.]Acknowledgements
Many people helped to make the superprize possible. Thanks are due to Peter Niehenke for translating the superprize article into German, Dr David Nias for helpful comments, Jane De Rome for a summary of the Dean-Elwell debate and for researching past AA prizes, Austin Levy for helping with the control entry, the sponsors for putting up the money, the judges for their careful and frequently time-consuming assessments, and of course the entrants for such an interesting and diverse assortment of entries.
The rest of the article appeared in Astrological Journal 1986, 28, 92-96, 274-275; 1987, 29, 86-90, 143-147, and in FAA Journal 1986, 16(1), 65-72, as follows:
Superprize Results Part 2: Details of each entry
Entries judged (15 cases)
1. Is it clear what was done?
For point 5. the judges needed to know how the observed results compared with those expected by chance, or how the results for authentic charts compared with those for control (ie non-authentic) charts. This allows the astrological contribution to be assessed while keeping everything else constant, which is what the superprize is all about; to do otherwise is like expecting a prize for discovering that all males have Mars in their charts. Obviously if the entrant gives no expected frequencies or uses no controls (and most of them did not) then no judgement is possible. To show you what the judges had to cope with, and, hence the pitfalls facing any researcher, we have summarised the main problem areas for each entry. The verdict given for each entry indicates the points (if any) on which the judges were unanimous that the entry had failed, and includes a verbatim selection of judges' comments. This section has been checked by the judges, and then by the entrants, and any comments have been incorporated.
Barclay UK. 3 pages. A description of 2 horary charts that successfully predicted a satellite landing in the Indian Ocean 4 days in advance, and the location of a cat lost for 10 days. Problems: sample too small, no indication that sample is unbiassed. Verdict: fails 3, 4 and 5, coincidence must be shown to be unlikely as an explanation, satellite is not identified, further satellite predictions made weeks or months in advance would make this sort of evidence far more believable.
Beyse Germany. 40 pages. Description of parent-child contacts for 11 families. Contacts were 0, 45, 60, 90, 120, 135 and 180 aspects, orb 5, between Sun, Moon, Venus, Mars, Moon's Node, Ascendant and M C in one chart and those in the other. Inspection of the author's data showed that the average number of reported contacts was 7.4 (range 1-20) for 58 pairs of related charts vs 16 expected by chance, the difference being due to contacts present but not reported, hence the reported contacts can be entirely explained by chance. The author claims her results show (1) that astrology is true, (2) that each ecliptic de ee has a typical physical property, and (3) that form should be recognised as a dimension. Because astrological factors have many correspondences with nature, for example the elements represent combinations of heat and moisture, astrological symbols are a mental reality. Problems: no summary of observed frequencies (so the reader has to do all the work), no expected frequencies (which can of course be overcome by controls involving parallel studies of unrelated charts, but there are no controls either), so the entry is impossible to judge. Verdict: fails 3 and 5, interesting study but lacks statistical evidence.
Brummund Germany. 9 pages. Comparisons of 50 famous artists born 1890-1954 with 34 mongoloids born 1966-1974 using the Uranian system of 10 planets, 8 hypothetical planets, Ascendant, MC, Moon's node, and 0° Aries. The author counted the number of contacts between A and B using multiple of 22.5-degree aspects and 10 orbs, where A = a significant point such as the Sun, 0° Aries, or the Hades/Admetos midpoint, and B = the remaining points, all midpoints, and antiscions of these points and midpoints. The hypothesis was that artists would have more contacts than mongoloids. In fact they had less, the mean per chart per significant point being 44 for artists and 47 for mongoloids vs 45 expected by chance. The contacts showing the biggest difference between artists and mongoloids for the three significant points for which the author gives results are as follows:
A = Sun A = Moon A = Hades/Admetos
The above figures are the mean number of contacts per chart. The author claims that only the hypothetical planets differentiate between artists and mongoloids, but this is not evident from the above results, where hypothetical planets are not noticeably more involved than authentic ones. Furthermore it is not clear whether any differences are real or merely due to chance (aggravated by the large number of factors tested) and to differences in the span of birth years. Verdict: fails 4 and 5, samples too small, no replication, results negative.
Ferguson UK. Two entries. (1) Can astrologers predict the chart factors that relate to outstanding success in dancing? Entry consisted of the first chapter from a 250-page book Stars of Dance, plus notes specially prepared for the superprize, analysing the birth charts of 1043 ballet dancers and 510 non-dancers, all with birth times. Data were obtained by the immense labour of writing to 6000 dancers direct. Twelve famous dancers or dance directors each nominated 6 traits required for outstanding success in dancing. The results condensed to 20 traits clustered around determination and artistic sense. The chart factors corresponding to these traits were then nominated by each of three professional astrologers to give a total of 60 factors. Computer analysis of several thousand factors for dancers vs non-dancers confirmed 6 of these factors (Moon-Neptune, Mercury-Saturn and Mars-Jupiter aspects, Venus and Jupiter in Scorpio, Saturn seldom on IC) and 5 other factors not selected by the astrologers (Mercury-Uranus and Venus-Saturn aspects, Sun and Mars on IC, Uranus on MC). Problem: no indication of the match expected by chance, so the results are impossible to judge. (2) Does the birth chart show sex differences? Entry consisted of the Stars of Dance results for 147 male non-dancers and 313 female non-dancers that showed astrological differences between males and females. Problem: similar studies have found no evidence of sex differences, the most convincing differences were said to lie in the distribution of ascending signs but by chi-squared test the differences are exactly at chance level (p=.5 df=ll). Verdict: (1) fails 3 and 5, (2) fails 5, one of the better entries, no statistical tests for the entire data set, the reported differences are not convincing. This entry won a consolation prize.
Glotz Germany. 15 pages and 2 charts. The claim that the charts of successful, talented people show a harmonious pattern (ie an above average number of contacts between Jupiter and the Sun, Moon, Mercury or Venus), whereas those of aggressive, uncontrolled people, such as murderers, show a tense overall pattern (ie discordant aspects between Pluto and the Sun or Mars, between Saturn and Mars or Uranus, and no Grand Trine). Problems: no observed frequencies, no expected frequencies, no controls, no details of how the sample was selected, so it is impossible to judge. Verdict: fails 3 and 5, sample size is woefully inadequate, no account is taken of non-astrological factors, no indication that samples are unbiassed, study must involve blind assessment, examples no matter how good are not enough.
Two of Glotz's statements are worth quoting because they probably represent the feelings of most astrologers. (1)"The horoscope tells us about such features as stamina, intelligence, courage, personal dominance, adaptability, single-mindedness, as well as the way in which a person expresses himself, and specific talents, such as artistic and manual skills, eloquence, critical judgement, to name but a few. It also reveals the threats to which a person is exposed and the times at which these threats will present themselves." In other words almost everything can be explained by astrology. (2) "Studying astrology is an exciting adventure which never ceases to be interesting. Time and again it leads to surprising insights and astounding results. For astrology no evidence needs to be furnished. Astrology furnishes this itself day after day to all those interested who approach this science without bias." But phrenologists said the same about phrenology.
Hill USA. 19 pages. Mostly devoted to a statistical sun-sign study of 913 jazz musicians broken down by instrument into groups of 80-180, and 1153+1134 football players broken down by position into groups of 80-330. Author points out that previous studies have involved topics too broad to give reliable effects, hence her study involves narrow topics. Many apparently significant distributions were obtained. Problems: no demographic corrections and no allowance for the many pitfalls of sun-sign studies (eg the virtual impossibility of calculating expected frequencies due to unavailability of data), so it is difficult to judge what is happening. Verdict: fails 3, 4 and 5, many of the sample sizes are too small, author lacks understanding of inferential statistics and of the difficulties involved in sun sign studies. Apart from this the entry was a model of diligent industry, and showed the kind of spirited effort that more astrologers should be putting into their research. If only the topic and design had been better the entry would have won a prize.
Klein UK. 10 pages. Can partners in a relationship pick the authentic synastry description from control descriptions? 50 volunteer couples were recruited via an announcement at a TA psychologists' conference in Switzerland attended by 750 delegates. [TA = transactional analysis based on child, adult and parent states that are assumed to exist in all of us.] Each partner received one authentic synastry interpretation plus four others, and ranked them according to accuracy. Each interpretation was based on major aspects between charts of orb 3 degrees, and house placements of the other's planets, and averaged 60 statements of 10-20 words each. 36 responses were received, all but 4 being from couples where both birth times were known. The authentic interpretation was ranked 1, 2, 3, 4, 5 in 12, 5, 13, 4, 2 cases respectively, which is significantly in favour of the authentic interpretation being chosen. Problem: other than excluding all references to natal sign and house position, knowledge of astrology was not controlled. Among so many delegates it is likely that some would know their synastry, and their inclusion would have been favoured by the selection process and by the subsequent ranking task (which subjects found difficult). Such subjects would obviously tend to pick the authentic interpretation. So it is impossible to tell if the results mean anything. Verdict: fails 5, follow-up studies should be conducted. This entry won a consolation prize and is an excellent example of the kind of basic research that more astrologers should be doing.
Niehenke and Boer, Germany. 15 pages. Can astrology pick people who are accident prone? Studies have shown that people who cause serious accidents, eg traffic accidents tend, to be of low intelligence, childlike, egocentric, aggressive, antisocial and irresponsible. The authors point out that these traits correspond to the negative side of Mars and Uranus and therefore should be predicted by the chart. An independent institution provided birth data, accident data and personality data for 6 persons who had caused accidents, each with 1-4 controls of the same sex, the same age within 4 years, without. social problems and without an accident record. The authors scored 5 hits vs just under 2 expected by chance, which is significant at the p=0.03 level. Problems: the sample is too small to guard against coincidence. The chart factors are not described so their expectancy in the general population (and hence whether they are realistic or not) cannot be determined. Verdict: fails 3 and 5, sample too small, interesting but statistically not convincing. This entry won a consolation prize for its design.
Pauk Germany. 27 pages. Can transits (Uranian system) predict events? For a female subject born in 1924 with a birth time recorded to half hour, 20 predictions were made in advance for 1982 and 22 for 1983. At the end of the year the subject (who knew what the predictions were) described what actually happened. Typical predictions were: (1) educational questions, intellectual conflict with females; (2) furtherance of hopes and wishes; (3) compulsion to achieve; (4) possible accidents; (5) accomplishments and satisfying work in business; (6) new beginnings, relief work, great changes. Author claims the accuracy was 93% in each case. Problems: predictions were generally vague and were known by the subject in advance (the ideal self-fulfilling prophecy), no objective methods of scoring, no controls, probability of agreement is impossible to determine. Verdict: fails 3 and 5, blind judgements by neutral observers are necessary.
Platz Germany. 80 pages. Can astrology predict accidents? From an examination of 500 accidents taken from newspaper accounts giving time, date and place, the author concludes that certain transitting and directed contacts must exist for an accident to happen. His system involves 14 natal points including Transpluto, their midpoints, 4 galactic points, multiple of 7 aspects of orb 20', and 12 different rates of direction applied simultaneously. Problems: no expected frequencies, no controls and no statistical overview (essential when so many factors are involved), so it is impossible to judge. Verdict: fails 3 and 5, very interesting hypothesis but no convincing statistical confirmation.
Rupert USA. 11 pages. Can astrology predict 16PF scores? From the charts of 21 subjects aged 18-49 the author predicted their scores on the 16 scales of the 16PF personality inventory. The predictions were then compared with the actual scores obtained independently by a psychologist. Significant correlations were obtained for 8 traits, namely warmth, intelligence, rebelliousness, compulsivity, and self-sufficiency, for which p=0.01 corresponding to r=0.55 for df=19 (no actual correlations were reported), and boldness, suspiciousness, and shrewdness, for which p=.05 corresponding to r=.43. Non-significant results (r<0.30) were obtained for conformity, emotional sensitivity, anxiety, ego strength, dominance, impulsivity, imagination, and guilt proneness. Predicted and actual scores were given for one subject, said to be typical, for which the correlation was a remarkably high 0.80 (df=14, p=0.0002). The study was made in 1980. Problems: low validity of 16PF scales (hence the correlations appear to be higher than they should be even if astrology were perfect), no controls, no allowance for self-attribution, insufficient description of procedure and results (adequate reporting is especially important here because later better-controlled studies have consistently failed to confirm such findings). Despite these problems the entry is the kind of study that more astrologers should be doing. Verdict: fails 3 and 5, self-attribution could readily explain the results.
Schneider-Gauquelin France. 25 pages plus 70 pages of reference material. Do the Gauquelin findings re key sectors improve the accuracy of astrological interpretations? The point here is that key sectors, ie. cadent houses, are traditionally regarded as weak and therefore, tend not to be used, whereas according to the Gauquelin findings for eminent professional people the opposite should apply. So if the Gauquelin findings can be found (albeit in disguise) in traditional descriptions matching chart to personality, which implies that the Gauquelin findings have to be used to make the chart fit, and if the Gauquelin findings apply to ordinary people, then the accuracy of chart interpretations is indirectly supported. The author cites studies suggesting that the Gauquelin findings do apply to ordinary people. For example Steffert and Shanks found that happily married people tend to have the Moon in key sectors whereas unhappily married or divorced people tend to have Jupiter in key sectors. And Bollen found that in 5 generations of her Dutch family the son who inherited the farm was not the oldest (as was customary) but the one with Jupiter in key sectors. The author then selects the charts of ordinary people (10 cases) from Lois Rodden's Profiles of Women, a book of charts and personality descriptions published in 1979, and shows that when one of the five Gauquelin planets is mentioned in the interpretation (meaning that it is seen to be relevant) it tends to be in a key sector. For 28 mentions the observed frequency in key sectors was 17 vs 5 expected, which is significant at the p=0.002 level. Interestingly, the author shows that Rodden overcomes the problem of weakness due to cadent house placement by finding other reasons (eg close aspects) to mention the planet. Problems: small sample size, Rodden's descriptions may have been influenced by knowledge of the Gauquelin findings, the applicability of the Gauquelin findings to ordinary people is uncertain (negative studies are currently as common as positive studies). Verdict: fails 5, a lot of interesting material but is any of it really convincing? This entry won a consolation prize for its presentation and organisation of reference material.
Sohnius Germany. 11 pages. Does colour preference predict sun sign? The relation between sun sign and scores on the Pfister colour pyramid test was determined using 350 subjects aged 18 and over. The Pfister test involves arranging 24 coloured squares inside a triangle to give the most pleasing effect. A further 56 subjects divided into two equal groups were then given the Pfister test, and for each subject the 12 signs were ranked blind in order of best fit with the previous results.
The correct sign ranked 1-3, 4-6, 7-9, 10-12 in 10, 6, 4, 4 cases (first group) and 14, 5, 6, 3 cases (second group), showing a distinct trend in favour of the correct sign. After eliminating subjects who had ambiguous colour scores (ie over-emphasis on either colour or form) or who were born near a cusp (18th-25th of each month), the corresponding results were 10, 4, 0, 0 and 13, 2, 1, 0 respectively, showing a highly significant trend in favour of the correct sign (for the combined results p=10-9). Problems: No check on whether the preferred colours are those traditionally associated with each sign, in which case the results could be explained by self-attribution. An independent replication with 193 subjects made after the entry was submitted gave negative results. The poor validity of colour tests is of no consequence here because no interpretation is involved, but if the test has poor validity do the results mean anything? Verdict: striking findings, procedure not clear, deserves a follow-up. This entry won a consolation prize.
Starr USA. 50 pages. A description of a sign study of 193 cancer victims and 193 people who died from causes other than cancer. The planets Sun through Saturn and the Moon's North Node were tested. One cancer distribution (Mars, peaking in Cancer and Pisces) and one non-cancer distribution (Node, peaking in Gemini and Virgo-Libra) were significant at the p=0.05 level, total 2 vs 0.05 x 8 x 2 = 0.8 expected. Problems: no demographic corrections, expected frequencies are dubious eg those for Venus in each sign were taken incorrectly as 1/12. Verdict: fails 3 and 5, does a nice job of explaining what was done but has virtually no discussion of non-astrological factors, does not satisfy the requirements of a proper scientific study.
Urban-Lurain USA. 77-page MA thesis and book Astrology as Science -- A Statistical Approach (AFA 1984). Can astrology distinguish alcoholics from the general population? A multiple discriminant analysis was performed on 53 AA (Alcoholics Anonymous) members and 217 members of the general population matched to the same birth year range (1921-49) and sex ratio (60% male) as the AA sample. The chart factors were angular separations between planets, Moon's Node, angles (ie Ascendant and MC) and Part of Fortune excluding those between the outermost planets, total 96 pairs (exactly how the angular separations were used is not stated, discrete aspects and orbs were not used), plus the house position of planets, Node, Vertex and 8 parts, total 20 (house system was not stated). The six factors that best discriminated between alcoholics and controls, in descending order of importance, were angular separation Moon-Ascendant, house position of Venus, and the angular separations Pluto-Part of Fortune, Mars-Neptune, Sun-Venus and Sun-Saturn. When the discriminant function was used to classify the data from which it was obtained, it correctly classified 81% of the alcoholics and 81% of the controls (which is a typical result for this kind of analysis even if the data be random numbers). However it also correctly classified 72% of a second matched set of 230 controls, which is significant at the p=10-10 level and therefore suggestive that a genuine effect exists. Problems: (1) Of the 116 factors tested, 70 involve angles, parts or house position and are therefore sensitive to birth time errors. Yet half the alcoholics have birth times given only to the hour or half hour. So do the results mean anything? (2) The standard procedure in a multiple discriminant analysis is to divide the test cases and controls into two groups to see if the results replicate. But only the controls were divided, leaving any artifacts in the test cases. Without proper replication no conclusion is possible. (3) The statistical tests assume that the variables are independent. However, although 96 angular separations between 15 points were analysed, only 14 separations are needed to completely define the others. Similarly a part is completely defined by its components. So such variables are not independent. Since the statistical assumptions [of independence] are not met, do the results mean-anything? (4) In the USA alcoholism and heavy drinking in 1980 were about 5 times more prevalent in men than in women. Unless we pretend that charts also indicate sex, this means that even if charts are 100% accurate for men they can be only 20% accurate for women, and vice versa. On average the maximum possible accuracy is therefore (100+20)/2 or 60%. On this basis the observed accuracy of 81% and 72% is too good to be true. (5) The results conflict with those of the handful of other studies which found no evidence that the chart relates to alcoholism (See Dean G. Alcoholism revisited. Astrological Journal 27(3), 179-182, Summer 1985).
Verdict: the judges found the entry overloaded with jargon and difficult to follow. They agreed that it was a nicely conducted study but due to the above problems were unable to decide what the results indicated. A recent independent review by Francoise Schneider-Gauquelin, NCGR Research Journal 3(3), 45, September 1985, reached the same conclusion: "We miss in this work a clear view of what exactly has been evaluated. A mathematical model of this complexity, applied to a sample of only 53 data (which emphasizes the role of chance fluctuations), can produce results for various reasons. They may be astrological, but they may also be due to demographic, astronomical, etc, particularities of the target sample." This entry won a consolation prize.
A moral for research astrologers
In our opinion there was no excuse for faulty design because: (1) In most cases proper design would have involved no more effort than did the actual design, for example dividing a sample into two replicate groups is as easy as not dividing it. (2) Ample guidance exists via Recent Advances, via Correlation, and via almost any textbook of experimental psychology. (3) The standard required is no more than that expected of any contribution to Correlation.
The moral for astrologers interested in research is: make sure your design is right, otherwise your results will be worthless. We have often been asked how good research should be done, as if there was some magic formula that would make you an expert researcher overnight. Sadly, there is no such formula, anymore than there is a formula that would make you an expert astrologer overnight. For those who wish to do good research there is really no alternative but to become seriously interested in the subject, to read up the literature on experimental design, and to exchange ideas with those already doing good research.Browse through the selection in your local library and pick the one you fancy. If this produces nothing, try: S Miller, Experimental Design and Statistics, Methuen, London and New York, 2nd revised edition 1984, paperback, 186 pages, an unusually clear and readable account by a psychologist for absolute beginners. Contains explanations and summaries of how and why of design, step by step instructions, tables for most of the statistical tests likely to be needed, and worked examples, but no astrological topics (for which you will still need Recent Advances and Correlation).[Today the examples of research given on this website, especially in Research Results under Doing Scientific Research, are the best next thing to hands-on guidance.]
Appendix: The fake superprize entry
As mentioned in the text, the fake entry was designed to appear plausible to scientific judges despite being a complete lie. Results such as those supposedly reported below have never been observed.
A Correlation between Transits and Events
Abstract -- The few controlled studies of character interpretations from the astrological birth chart have produced negative results. There have been no controlled studies of event interpretations, and the aim of the present study was to rectify this deficiency. The astrological hypothesis is that planetary positions at the time of an important life event, e.g., marriage or the death of a parent, show a correlation with planetary positions at the time of birth.
The hypothesis was tested using 203 ordinary people typical of those who visit astrologers. For each subject the birth data and the dates of at least ter important life events were accurately known. The analysis involved all aspects that are multiples of 30 degrees, allowing zodiacal positions to be reduced to positions on a 30 degree scale, thus increasing the frequency per degree and consequently the reliability of the test. The orb was one degree. These aspects and orb are used by the majority of astrologers.
The subjects were divided at random into two replicate groups of 101 and 102 subjects respectively. For each subject the planetary positions at birth and at the time of each event were calculated by computer and converted to two frequency distributions across 30 degrees, viz. one for the birth chart and the other for the combined events. The extent of the association between the two distributions was then determined using the product-moment correlation coefficient. Control events were generated for each subject in three different ways, viz. by using the events for another subject, by using random dates for the same event years, and by using random dates and random years.
If events tend to occur at the time an aspect exists between an event planet and a birth planet, then the event planets should tend to cluster around the aspect points to the birth planet, consequently the correlation between the birth distribution and the event distribution should be positive and significant. The average observed correlation for the two replicate-groups was 0.56 (p = 0.001, df = 28) and 0.51 (p = 0.004, df = 28) respectively, whereas the correlations using the control events were all at chance level, the six means ranging from -0.01 to 0.20, average 0.11, with no significance level being less than 0.3.
Some of this correlation could be explained by astrological artifacts and by subjects planning certain events, e.g. the date of a marriage, according to astrology, but not by sampling bias, which was not detectable. However, when only those dates over which the subject had no control were used, e.g. the death of a parent, the average observed correlations of the two replicate groups were still significantly positive at 0.45 (p = 0.006) and 0.44 (p = 0.015) respectively. The effect is inexplicable and is evidence that the accuracy of chart interpretations (in this instance interpretations involving life events) cannot be explained by non-astrological factors.