From       110m 9g 210kb       Home       Fast-Find Index

Research results (91 abstracts)
From Correlation, APP, AinO, and Kosmos

Abstract -- Contains abstracts of 91 studies, most of them empirical, from four astrological research journals. There are 37 abstracts from Correlation: Journal of Research into Astrology 1981-2007 published by the British Astrological Association, 22 from the now defunct Astro-Psychological Problems 1982-1995 published by Francoise Gauquelin in France and (in 1989-1990) by the National Council for Geocosmic Research in the USA, 18 from Astrologie in Onderzoek [Astrology under Scrutiny] 1986-2003 published by Wout Heukelom in the Netherlands, including its precursors 1977-1985 published by NVWOA the Dutch Society for Scientific Research into Astrology, and 14 from Kosmos 1978-1994 published by ISAR, the USA-based International Society for Astrological Research. At the time the first three journals were the world's only peer-review astrological journals devoted to scientific research, whereas Kosmos was more an astrological journal than a scientific research journal, hence the fewer abstracts. The abstracts are comprehensive, averaging 270 words (range 80 to 950), and are annotated with later information where necessary. Most are from 1980-2000 when scientific research into astrology was at its peak. There are nine figures. These abstracts are the next best thing to being there. They illustrate the topics then being investigated by astrologers and others, the immense labour that could be involved, the results that were invariably incommensurate with astrological claims, and the then intense scientific interest in astrology that (in view of the negative results) will most likely never arise again. The areas covered are: Character (11 abstracts), Events (17), Signs (16), Aspects (5), Houses (3), Gauquelin Effect (11), Tests of Astrologers (12), Approaches to Research (16).

The following abstracts are limited to the journals Correlation, APP, AinO, and Kosmos and therefore do not cover research reported in orthodox journals or in the few other astrological research journals such as the French Les Cahiers du RAMS founded 1993 (abstracts in English are at and the American Federation of Astrologers Journal of Research founded 1982 (although by 1995 only seven volumes had been published). Nevertheless they illustrate the wide range of topics being investigated by astrologers and others. Overviews of all studies (not just those abstracted here) relevant to Sun sign self-attribution, tests of astrologers and clients, lunar effects, and Gauquelin's tests of signs, aspects and planets, are given in Tests on this website under Tests of Astrology.

The areas covered are:
1 Character 11 abstracts eg extraversion, harmonics, marriage, murder, Pluto, redheads, terrorism.
2 Events 17 abstracts eg accidents, death, elections, Moon planting, radio, rain, wars, weather.
3 Signs 16 abstracts eg Age of AQ, guessing, patterns, ray theory, self-attribution, sidereal.
4 Aspects 5 abstracts eg graphs of non-uniform occurrence, occurrence in eminent people, validity.
5 Houses 3 abstracts covering house division, rulerships, tests of validity.
6 Gauquelin effect 11 abstracts eg interpretation, replication, unsolved problems, 8000 tests of.
7 Tests of astrologers 12 abstracts eg case histories, death, extraversion, picking own reading.
8 Approaches to research 16 abstracts eg computers, internet, judgement bias, need for science.

Within each category the abstracts are in date order. Where the original abstracts are inadequate or non-existent (which is most of them) they have been expanded, often very considerably. Average length is 270 words, range 80 to 950. Occasional notes in [ ] provide later information where relevant such as failure to replicate. The source material is shown in the picture below.

Source material

Source material. Top: APP 1982-1995 including APP-NCGR, and AinO and precursors 1977-2003. Bottom: Correlation 1981-2007 and Kosmos 1978-1995. Altogether about 4.6 million words (Correlation 37%, APP 13%, AinO 30%, Kosmos 20%) occupying 90 cm of shelf space.

Latest updates
In July 2013 the best of this archive was updated, expanded, and put into book form. The new book contains more than a hundred abstracts of tests that are not covered below.

Click here for details.

Rise and fall of interest in research
Scientific research into astrology became popular during the 1980s due to the advent of personal computers in the late 1970s, which removed the calculation barrier, and the publication in 1977 of Recent Advances in Natal Astrology, the first fully-referenced critical review of empirical findings. Within five years there were the three peer-review journals AinO, Correlation and APP that subsequently led the field. But all was not well. Results were invariably incommensurate with astrological claims, and since the 1990s interest has slowly declined, see the plot below. The period's intense scientific interest in astrology will most likely never arise again.

Combined word counts peak in the 1980s and 1990s
Combined word counts for the three peer-review journals plotted every five years. Total is about 3.7 million words, of which 46% is from Correlation, 16% from APP, and 38% from AinO and its precursors. Some of the decrease in 1996-2000 is due to the demise of APP in 1995.

1. Astrology and Character (11 abstracts, see also Signs)

Research on Astrological Factors Between Married Couples
Thomas Shanks APP 1983, 2.1, 12-16. Reprinted in Kosmos 1985, 14.1, 16-26 but wrongly attributed to Marie Scheider, see 14.2, 14. In 1952 Carl Jung compared astrological factors between 483 married couples divided into three groups. There was a tendency to favour Moon conjunct Moon, which was in agreement with tradition, but it did not replicate, and the underlying statistics (which assumed uniform distributions) were inadequate. The author therefore repeated Jung's experiment with 960 married couples chosen at random from Gauquelin's first heredity experiment, and compared the observed frequencies of conjunctions and oppositions (orb 10 degrees) between all possible pairings of Sun, Moon, Ascendant and MC, with those for the same couples in 199 sets of randomised pairings. According to tradition these factors should be prominent between married couples. But none reached the same prominence as in Jung's results, and none were even marginally significant. There was a fairly uniform distribution of significance levels across the entire range from 0 to 1, which is precisely what would be expected if there were no astrological effects. Conclusion: the frequency of these factors does not differ between married couples and unmarried randomly matched couples.

A Computer Study of Relationships
Gail Guttman Kosmos 1985, 14.2, 2-8. Relationship astrology was tested on 106 long-term couples and 145 short-term couples (lasting less than ten years) by a computer program that tallied 17 planets and asteroids, 4 angles (Ascendant, MC, Anti-vertex, East Point), and 16 aspects (all multiples of 30 and 45 degrees, orb 1-3 degrees), making a total of 3360 possible factors. The same couples were mixed and rematched to form a control group. In decreasing order of statistical significance the top seven male-female contacts for the long-term couples were Moon-Moon's mean south node, East Point-Ascendant, Chiron-Saturn, Moon-Moon's true north node, Pluto-MC, Ascendant-Pluto, and Ascendant-Juno. For the short-term couples they were Moon's mean south node-Uranus, Pluto-Chiron, MC-Mars, Ceres-Vesta, Jupiter-Antivertex, Chiron-Pluto, and Neptune-East Point. Of the top 15 contacts none replicated between the two groups, and traditional contacts such as Sun-Moon and Venus-Mars were absent. Obviously there is much here that needs further study. [The calculation of significance is unclear. Since none of the top 15 (somewhat bizarre) contacts replicated, and none were predicted in advance of testing, the results provide no evidence for astrological effects in relationships. All that can be concluded is that in any list of ranked data, even random data, some will come top.]

Sixty-Eight Birth Charts of Terrorists
Luigi Squitieri APP 1985, 3.2, 16-21. Definite statistical conclusions can hardly be drawn from a study of 68 cases only. But the author's clear starting hypotheses (eg low frequency of Cancer and 4th house, high frequency of 11th and 12th houses, angular Mars, Uranus and Neptune) are so strongly contradicted by the outcomes that more cases would probably not change the present conclusions: tradition is not confirmed in the case of terrorism. Includes birth data.

Can Astrology Predict E and N? 1: Individual factors
Geoffrey Dean Correlation 1985, 5.1, 3-17. To test whether astrology can predict E (extraversion) and N (emotionality) in ordinary people, subjects with extreme scores on the Eysenck Personality Inventory (EPI) were selected from 1198 subjects mostly from the southern hemisphere, all of whom had known birth times. The result was two replicate sets each with 54 extreme subjects for each of E+, E-, N+ and N-. The average pair of opposite extremes was roughly equivalent to the two most extreme persons in a random sample of 15 adults. The following factors were tested: tropical signs, decans, elements, sidereal signs, aspects, harmonics, hemisphere, Gauquelin plus zones and angularity, both individually and in combination. The latter excluded sidereal signs but included midpoints and Placidus houses. If astrology predicts that factor X indicates, say, E+, then the frequency of X should be higher in extreme E+ subjects than in extreme E- subjects. The results of 132 such tests, and a multiple discriminant analysis, showed that no factor performed consistently above chance level.

The Meaning of Pluto Part 1: Experiment and primary analysis
Terry Dwyer Correlation 1987, 7.2, 9-13. A sample of 175 mixed adults completed a 36-item questionnaire addressed to likely Pluto meanings, namely intensification, exaggeration, social power, transformation, and suppression. Answers were on a 7-point scale from 0 disagree to 6 agree. A comparison with the natal charts (divided into those with the relevant Pluto aspects and those without) found no evidence to support any of these meanings. Combining into hard aspects vs soft, or close aspects vs wide, or weighting by aspect strength, produced no improvement. The results were suggestive of no meaningful trend whatever, which does not deny that Pluto may have meanings other than those t4sted.

The Meaning of Pluto Part 2: Further analyses of Dwyer's data
Geoffrey Dean and Rudolf Smit Correlation 1987, 7.2, 14-21. Dwyer's study is one of a mere handful of attempts in the history of astrology to investigate a planetary meaning by direct experiment. Despite his negative results, his sample size was sufficient to detect any useful Pluto effect, provided of course that Pluto's real meanings are among those addressed by his questionnaire. On the other hand, it is possible that his negative results arose because single questions are generally poor measures of what they are supposed to measure, which can be remedied by adding together the answers to questions that relate to the same thing. Our further analysis puts this point to the test.

Factor analysis of the correlations between responses to Dwyer's 30 questions revealed three underlying themes, namely Power (intense, needs to achieve), Suppression (reserved, hide's feelings), and Transformation (does own thing, prone to changes and upheavals). The three themes, each based on the seven questions most strongly correlated with them, were then compared with Pluto strength measured in three ways, namely (1) strength by aspect weighted by planet and orb, for aspects that were multiples of 45, 30 and 7.5 degrees, (2) strength by Pluto angularity based on orb, and (3) a measure obtained by multiplying (1) by (2), which approximates what an astrologer does when interpreting a chart. The number of significant correlations between theme and Pluto strength were very close to those expected by chance. No aspect multiple was better than another. Approach (3) showed no improvement over (1) or (2), nor did charts with birth times (N=121) over those without. Results using random numbers were just as good as those using astrology. There was no hint that Pluto by aspect or angularity means Power, Suppression or Transformation. Our improved approach was more than adequate to detect a worthwhile effect but it merely confirmed Dwyer's negative findings. In the hope that others could do better, we offer all data and computer programs free of charge to anyone who is interested. [During twenty years there have been no takers]

More about Uranus and Murderers
Lorna Houston, Francoise Gauquelin, Arno Muller, and Mike O'Neill APP 1988, 6.3, 23-25. Lorna Houston having obtained a significant frequency of Uranus retrograde positions in the birth charts of 132 murderers (see APP 6.2) she checks on a new sample of 491 murderers to see whether this result is replicable. Her new result is not significant but is in the same direction. So does it confirm the first outcome? Two readers of APP (Muller and O'Neill), one of them a university professor, tell us how astronomical and demographic effects could explain the outcomes. This will have to be further investigated. [Follow-up checks on the original data appeared in APP-NCGR 1989, 7.2, 27-31 between Pottenger, Muller and O'Neill, but demographic effects proved to be too strong to allow a clear conclusion. Inexplicably an accompanying study 23-25 by Lorna Houston using a new sample of 110 murderers from the Canadian Centre for Justice Statistics did not report the occurrence of retrograde Uranus.]

Mercury stations
Frank Jakubowsky Kosmos 1994, 23.1, 38-42. Mercury retrograde has a bad reputation among astrologers, being held to indicate weakness or confusion in areas ruled by Mercury. To test this idea all entries in Who's Who in US Writers, Editors & Poets 1986-1987 with birth dates in 1900-1949 were analysed, a total of 4997 births. Mercury was retrograde (including the subsequent station) for 1000 births vs 955 expected, a surplus of about 5%, which indicates that, contrary to its reputation, a natal retrograde Mercury is an advantage for writers. However, the distribution of births was uneven over the 50-year period, so the theoretical expectancies could be wrong. The study needs to be replicated. [By chi-squared test the surplus is not significant, p = 0.11]

The Mars-Redhead Dilemma
Nick Kollerstrom Correlation 1997, 16.1, 19-21. An early study found that redheads tend to be born with Mars within 30 degrees of the natal ascendant at above chance level. Did the Mars-redhead effect replicate? Much controversy was generated by this highly original project, coordinated by Hill & Thompson in the USA. The initial pilot study was gathered in March 1987, then by July of that year another 400 US redheads had been gathered to test the hypothesis. More recently a further 479 were gathered in the USA and Canada, plus 373 UK redheads. In addition there were 100 UK redheads gathered long ago by John Addey, where the birth data are lost and only the Mars equal-house frequencies remain. The results seem to show what could be called beginner's luck, with the percentage excess of Mars in key sectors decreasing with each subsequent sample. There was also a suggestion of a Mars-setting deficit, but this did not show up too well in the replication either.

John Addey's Dream: Planetary Harmonics and Character Traits
Geoffrey Dean Correlation 1997, 16.2, 10-39. According to John Addey, harmonic analysis of the Gauquelin trait data reveals the existence of planetary harmonics in the diurnal circle beyond the simple emphasis on key sectors. For example, consider Mars as it moves around the diurnal circle. If the full circle = 12 sectors, then the 3rd harmonic interval = 12/3 or 4 consecutive sectors, and the 4th harmonic interval = 12/4 or 3 consecutive sectors. Comparison with the Gauquelin trait data for sports champions seemed to show that those born with Mars in 3rd harmonic intervals tended to be modest while those born with Mars in 4th harmonic intervals tended to be aggressive. Higher harmonics seemed just as meaningful, for example sports champions born with Mars in 15th harmonic intervals tended to be steadfast. Such results became central to Addey's dream of a unified astrology based on harmonics. Earlier findings had so convinced him that in Astrology Reborn (1971) he was able to make his now-famous quote about the future of astrology: "From being an outcast from the fraternity of science, it seems destined to assume an almost central role in scientific thought ... its impact will be felt in the next twenty years" (pages 3, 23).

But both Addey's data and procedures are problematic. First, the trait data were not extracted blind and can be shown to contain bias due to Gauquelin's knowledge of planetary positions during the extraction process. Second, Gauquelin's own statistical tests were faulty because they were based on the number of traits when they should have been based on the number of subjects, leading to conclusions that have led everyone astray. Third, the samples were small sub-samples for which Gauquelin had extracted trait data, not his original large samples of births, yet the sampling requirements in harmonic analysis are more stringent than in other types of analysis. My computer simulations show that Addey's sample sizes were too small for his results to be meaningful. That is, the sampling errors across the 100 sectors Addey used for extracting harmonics 1-20 were always high enough to make the obtained harmonics unlike those in the original population. (My analysis simulated Addey taking a sample of births and then analysing their distribution to see what harmonics it contained, but unlike the samples available to him, mine could be of any size and the correct answer was known in advance.) The figure below shows the correlation between synthetic populations containing harmonics 1-20 of equal amplitude and random phase angles, and samples of different sizes picked at random and distributed over 100 sectors. The correlation is a measure of the sample's adequacy for harmonic analysis and should exceed about 0.7 if the results are to be meaningful. Each point is the mean of 100,000 replications.

Correlation between population and sample vs sample size
The correlations between population and sample attained by Addey's sample sizes (median = 65) are mostly negligible, showing that sampling errors have made Addey's samples too unlike the population to allow a meaningful harmonic analysis. If the number of sectors is reduced from 100 to 36, the sampling errors per sector are reduced, which increases the correlations as shown by the dotted line. Approximately the same increase is obtained if the harmonics extracted are reduced from 1-20 to 1-10.

Other weaknesses are ineffective criteria for selecting amplitudes (which led to amplitudes that were mostly noise being accepted as meaningful), non-independence of traits, sector bias (which created spurious odd-numbered harmonics), incorrect expectancies, non-uniform expectancies, and subjective follow up. Collectively these weaknesses are fatal. Despite Addey's inspiring vision and astonishing labour, it seems that most of his results can reasonably be attributed to artifacts and the rest to Gauquelin bias in trait extraction. The sad but safest conclusion is that planetary harmonics do not exist beyond the emphasis on key sectors. The same procedural weaknesses apply to Addey's harmonic work in general, which leaves his harmonic theory of astrology with no secure basis. His dream remains so far only a dream.

Meaningful Coincidences: Parallels between Phrenology and Astrology
Geoffrey Dean Correlation 1998, 17.1, 9-40. Phrenology (a system of reading character from brain development as shown by head shape) is now effectively dead but in the 1830s it was more popular than astrology is today. The story of phrenology is rich in lessons for astrology. But its literature is so huge, so clogged with side issues (of philosophy, of politics, of religion, of morality, of society in general), so often tedious to read (wordiness being the style of the day), and so difficult to find except in specialised libraries, that these lessons have gone largely unrecognised. Like astrology, phrenology encourages you to assess yourself and act on its findings to achieve harmony with the world. Like astrology, it flourished because practitioners and clients saw that it worked. It was claimed to be "so plainly demonstrated that the non-acceptance of phrenology is next to impossible". But the experience-based claims of phrenologists were completely wrong. We now know that a certain head shape cannot possibly mean what it is supposed to mean, even though the underlying philosophy of "know thyself" has undeniable appeal. Millions of people, unaware of the many ways their judgements could be led astray, had been seduced by phrenology's appeal into believing untruths. Could the same apply to the experience-based claims of astrologers? To answer this question I look at phrenology's social context, history, literature, testimonials, stock objections, and experimental tests, all of which have parallels in astrology. Until astrologers can demonstrate otherwise, the answer would seem to be yes. Whether the price paid for believing untruths is worth whatever satisfaction it brings, including keeping people from worse mischiefs, is a topic we might like to ponder.

2. Astrology and Events (17 abstracts, see also Signs)

A Lunar Sidereal Rhythm in Crop Yield and its Phasing
Nicholas Kollerstrom Correlation 1981, 1.1, 44-53. All Moon-gardening calendars agree that some kind of Moon sign effect at the time of sowing is important for maximising yield, but they do not agree on which zodiac should be used (tropical or sidereal) or on the zodiac signs which produce the best effect. Here experimental work will be used to examine these different possibilities. The crop yields vs Moon's position at sowing in experiments performed by Colin Bishop in Wales in 1976, 1977 and 1978 show that 27.3 day sidereal rhythms may be present. A wave-harmonic approach to such rhythms in three years of sowing data showed that lettuce maximum yield occurred in sidereal Water signs, and radish maximum yield occurred in sidereal Earth signs, the mean increase in yield being around 30-40%. Moon phase had no effect. The results can be criticised on account of the consistent small yields due to poor soil and sometimes drought conditions, and it is hoped that similar experiments will be performed under better conditions.

Bradley's Jupiter Pluvius Rainfall Study
Colin James III Correlation 1981, 1.2, 19-23. One of Jupiter's classical names is Jupiter Pluvius (Jupiter the Rainmaker). In the late 1950s the US astrologer Donald Bradley (1925-1974) found that heavy rainfall tended to occur when Jupiter was conjunct, square or opposite the local meridian at the moment of the Moon's entry into sidereal Capricorn prior to the rainfall. Bradley suggested that this supported an effect of both Jupiter Pluvius and a sidereal zodiac. Other than the work of John Nelson in radio disturbance [see next abstract], only Bradley's study provides hard geophysical support for an astrological theory. Previously unpublished data has aided my analysis of Bradley's Jupiter-rainfall study, and the results show that Bradley's findings are spurious due to the Moon's sidereal period being nearly commensurable with the Earth's rotation. The average time between the Moon's successive entries into sidereal Capricorn is 27.32 days, during which time the Earth rotates 27 full circles plus 89 degrees, while Jupiter moves about 2 degrees. The difference is close to 90 degrees, and it is this, compounded by Bradley's use of a wide moving total, that explains his findings. Consequently they provide no support for either a connection between Jupiter and rainfall or the existence of sidereal effects.

Shortwave Radio Propagation: Analysis of the Forecasts of John Nelson
Geoffrey Dean Correlation 1983, 3.1, 4-37. This study examines Nelson's claims that heliocentric planetary aspects correlate with shortwave radio quality (hard aspects make it worse) and that they can be used to improve the accuracy of forecasts. Computer analysis of 2006 half-day or quarter-day quality forecasts (based on planetary positions) made by Nelson for RCA during 1964-5, and 4960 daily forecasts made for 73 Magazine during 1966-82, failed to find support for his claims. There was no significant correlation between forecast and outcome (mean r = 0.01), and the outcome on days forecast as poor was not significantly different from that on days forecast as good. Nelson's forecasts performed considerably worse than US Government forecasts and a control forecast based on the quality one solar rotation before. The accuracy of 105 forecasts of solar flares was not significantly better than chance. To increase the sensitivity of the analysis a daily planetary index based on Nelson's rules was compared with observed radio qualities and geomagnetic indices. No planetary effect was detectable, nor was the alleged effect of nodes and perihelia. In disagreement with Nelson's claims, hard aspects and associated harmonic aspects were not consistently more numerous on the most disturbed 3% of days during 1969-80 than on the least disturbed 3% of days. In particular the 12 days with the most adverse planetary configurations during 25 years were not significantly different from those with the least adverse. Nelson's claims are incompatible with the physical processes involved and are shown to rest upon three things: (1) A statistical artifact due to the close but unequal spacing of aspect days, which means that small differences from radio days are more likely than large differences. (2) A calculation artifact due to counting forecasts as hits if they are within one unit of the observed quality, yet around 90% of all observed qualities fall within a range of one unit, so a hit rate of 90% (his claimed accuracy rate) is unremarkable. (3) Selection of data to fit the case. The results do not deny that the planets could affect the Sun in other ways.

Planetary Motions and the Occurrence of Earthquakes
Scott G Vail Kosmos 1985, 14.1, 2-10. Charts of the largest earthquake disasters seem to support every conceivable astrological factor yet no two are much alike. So it is obvious that statistical methods are necessary. The charts of 238 major earthquakes of magnitude 7.5 or more during 1930-1980 revealed no particular sign, house, or aspect preferences except (1) Saturn tended to cluster around the MC (52 cases vs 39 expected), and (2) for heliocentric planetary pairs excluding Uranus-Pluto there were 9% more applying (getting closer to exactness) multiple-of-45-degree aspects of orb 4 degrees than expected (assuming uniform planetary motion) and 8% fewer separating aspects than expected. These analyses were not subjected to statistical tests nor were there any controls, so the results are tentative. In any case, they do not indicate where earthquakes may occur. [No hypotheses were framed in advance, so the results may illustrate only after-the-event selection from many possibilities.]

Election to Government Ministries and Jupiter Transits
Grazia Bordoni, Ciro Discepolo, Vincent Grilli APP 1986, 4.3, 22-23. Statistical investigation of 834 elections of politicians to the Italian Government were carefully conducted by a group of astrologers from the South of Italy. Bias was avoided by taking all the representatives of all the 44 successive governments of the Italian Republic from the date of its creation to the date of our study. The hypothesis (consistent with tradition) was that there would be more transits (conjunctions and trines) by Jupiter to the natal Sun, Moon, Ascendant and MC on the day of their election than in a control sample of the same subjects with election dates randomly generated within three years (before or after) of the actual election date. The average frequency of Jupiter transits was slightly higher in the experimental group (0.399 vs 0.355) but the difference by t-test was not significant (p = 0.14). We had also planned to test transits by Venus, Saturn, Uranus, Neptune and Pluto, but the disappointing results for Jupiter (traditionally the most relevant planet) did not encourage us to do so. They require too much work.

Transits and Abduction of Children
David Peter Valentiner 1987, APP 5.3, 27-31. The author notes that empirical studies of transits have been generally negative. For example Ulrich Mees in APP 2.3 examined Sun and Moon transits (including the first twenty harmonic aspects) to natal Sun and Moon in 3045 deaths but found nothing significant. Nevertheless transits hold a prominent place in astrology. So the author attempts a more thorough conclusion by examining all hard transits (conjunction, square, opposition, orb one degree) between Mars and Saturn for 431 cases of child abduction supplied by the US National Center for Missing and Exploited Children. According to various textbooks, Mars-Saturn symbolism is consistent with abduction. For example in her 1976 book Saturn: A New Look at an Old Devil Liz Greene says "Actual physical maltreatment in childhood sometimes occurs with Mars-Saturn aspects" (page 116). The hypothesis was that Mars-Saturn and Saturn-Mars transits would be found more frequently than expected in the charts of these children on the day they were last seen. Expected frequencies were calculated by mixing each birth date with the abduction date of the next 100 records. The observed total (25 occurrences) was slightly higher than the expected total (19.3 occurrences), but not significantly so. The remaining 406 abductions (94%) occurred despite having no transit. The most significant transit was Mars conjunct Saturn, but even this was only marginally significant (p=0.10). Conclusion: the hypothesis is not supported. Alternative explanations, such as astronomical biasses or chance fluctuations, cannot be safely dismissed.

Traffic Accident Victims and Pluto squares
Francoise Gauquelin APP 1988, 6.1, 26-27. In a 1985 APP article Dr Kuypers studied a sample of 72 traffic accident victimes and found a general excess of squares in keeping with tradition. In particular 11 had Ascendant-Pluto squares (orb 7.5 degrees) vs 6 expected (p = 0.05). But some of the cases had been taken from the astrological literature, where selective publication in favour of tradition is normal, thus introducing a positive bias. A check against a new sample of 30 traffic accident victims taken mainly from newspapers did not show the same general excess of squares, but Ascendant-Pluto was still high (6 vs 2.5 expected, almost significant). A computerised check of the expectancies showed that they were slightly low, which reduced the observed significance, but Dr Kuypers can still argue that Ascendant-Pluto squares have held their promise. But for a skeptic, with the failure of other squares to give replicable results, and the precarious significance, the whole outcome has no convincing power that accidents and squares in the natal chart have any connection.

Horary Astrology is Unreliable
Martin Boot AinO 1989, 3.2, 2-12. In AinO 2.2 the author showed that, when applied by precise textbook rules, precise predictions by horary astrology are not possible. In AinO 3.2 Corry Sietsma argued that Boot was mistaken. She reads in his two example charts the exact events in question. In the present article Sietsma's reasoning is rebutted. She is not correctly applying the rules of horary astrology, which is why she can read anything she likes in the two charts. Thus Sietsma joins the multitude of astrologers who afterwards fit the chart to the event. [For a follow-up see the 1996 abstract under Tests of Astrologers]

Planetary Aspects at 500 Violent Deaths
Geert Thomassen and H.J.van Roekel AinO 1990, 5.2, 24-29, with a critique by Bert Terpstra AinO 1991, 6.1, 22-28. This abstract combines both. Thomassen collected birth and death dates from newspapers and graveyards for 500 persons who had died a violent death, usually in a traffic accident. By computer Van Roekel counted aspects between natal (at noon) and transitting (at death) planets except Moon. Harmonious aspects were defined as 30, 60, 120, 180 degrees, inharmonious as 0, 45, 90, 150 degrees. (Tradition would swap 0 and 180, but the allocation was based on earlier research; 135 was omitted to equalise the expected aspect numbers.) The total number of observed aspects was as follows:

 0-1     1-2     2-3    Orb 
1291 1285 1279 Inharmonious aspects
1170 1265 1243 Harmonious aspects
121 20 36 Difference

The differences are very small but in the right direction. However, contrary to tradition, the numbers do not decrease with increasing orb, and the outer planets Jupiter through Pluto showed the least effect. Conclusion: the outcome is unclear, but if there is an effect it is very small. The study needs to be repeated with clearer hypotheses, but it will not be easy to collect another sample in the short term.

1500 cases of death
Geert Thomassen and H.J.van Roekel AinO 1991, 6.1, 58. Thomassen visited several cemeteries and recorded a total of 1500 birth dates and their death dates. By computer Van Roekel counted aspects as in the previous abstract. According to tradition there should be significantly more inharmonious aspects at death than harmonious aspects. For three replicate groups of N=500 the transitting planets Jupiter through Pluto showed no clear effect. The rest (Sun, Mercury, Venus, Mars) showed a small excess of harmonious vs inharmonious aspects (average 721 vs 681), which was in the opposite direction to that predicted even though omitting Jupiter through Pluto after their results were known should have inflated any effect. The counts were also little different from the counts for three control groups of the same birth dates with random death dates (average 722 vs 706). Increasing the orb to two degrees did not affect the outcome. Conclusion: contrary to traditional claims there is no significant correlation between transits and death.

Birth Time Reconstruction
Bert Terpstra AinO 1992, 7.1, 7-9 and 1991, 6.1, 55-58 and 1990, 5.2, 20-23 and 1990, 5.1, 19-29. [This study of chart rectification took the author, an expert computer programmer, several years to complete and is perhape the most meticulous ever conducted.] In four articles the author describes a computerised method for reconstructing the correct birth time based on the traditional claim that no important event in life occurs unless a progression involving a chart angle (ie Ascendant or MC) is in force at the time. It is hard to imagine that a newly born child could record the movement of planets and angles in the hours after birth and then replay them in a way that would indicate external events. Nevertheless if progressions do work then the implications are tremendous, perhaps requiring the acceptance of analogy as an active principle, and it would pose a huge challenge to science. Not all primary and secondary progressions can be used in birth time reconstruction, only those that coincide with an event for one birth time and not another. This is true for primary progressions of the angles to natal planets, and for secondary progressions of planets to natal angles.

The starting point is (1) the subject's birth chart, (2) birth times at intervals of two minutes during say a quarter hour either side of the nominal birth time, and (3) a number of major events such as accidents. For the first event and each birth time the natal planets and angles are progressed at the rate to be tested (eg 1 day = 1 year in life, as here), and the date corresponding to each exact aspect to or from the angles is calculated for each birthtime. Contacts between angles are not counted. Within each 2-minute interval the number of hits between aspects and event are counted. To be a hit the aspect has to involve the right angle (Ascendant for physical events such as accidents, MC for social events such as marriage), the right aspect (hard for difficult events such as accidents, soft for easy events such as business success), and the right planet as defined by tradition (eg Jupiter for a promotion). If the same aspects fits several events occurring close together, only one hit is counted to avoid biassing the results -- the aim is to count aspects that match events, not events that match aspects. The small size of the interval (only two minutes) requires very high accuracy in the computed positions, which were computed with full astronomical accuracy, ie within 0.1 second of arc for the lights, angles and inner planets, and within 1 second of arc for the rest.

Finally the number of hits are plotted as a histogram vs birth time. Peaks in this plot identify the times that provide the best fit between chart and events. According to the traditional claim there should be a single pronounced peak at the true astrological birth time that accounts for all significant events. But this was never observed. For example, in his book on birth time reconstruction Horoscoopcorrectie in der praktijk (Ankh-Hermes, Deventer, 1985), the eminent Dutch astrologer Jack Chandu gives a detailed list of 179 dated events in his life, of which 17 are found in a pronounced peak in the histogram (see plot below), but not at the time of the first cry recorded to the second by Chandu's astrologer father using a watch synchronised to the radio time signal. (Chandu writes "I possess twelve birth times of myself, corrected to the second by well-known astrologers, none of which agree with each other. They [also] deviate many minutes from my actual birth time.") [Such disagreement is usual. For example various astrologers have worked backwards from the life events of Ronald Reagan to produce over 30 different birth times spanning 15 hours, each one said to be accurate by the astrologer concerned.]

Also, if progressions indicate events, important events should feature more prominently than unimportant events. To allow the author to make this check, Chandu selected 41 of his events as being the most important. But no peak contained all 41 events even when the method was applied to all 24 hours of the birth date, so the traditional claim (that no important event in life occurs unless a progression involving a chart angle is in force at the time) is wrong. Also only 4 of the 17 events in the peak were important vs 17 x 41/179 = 3.9 expected by chance, which implies that the peak is meaningless because it does not reflect importance. So there is no reason to assume it has anything to do with Chandu's birth time.

Furthermore, very little bias is needed in the selection of events to produce a peak in the histogram. For example if Chandu had included only 12 events in his list of 179 just because they fitted his progressions, it would have been enough to raise the average peak height of 5 hits (measured over 24 hours) to the 17 observed. So as a check the method was applied to the life of the author Graham Greene, for which 251 accurately dated events are listed in his biography The Life of Graham Greene by Norman Sherry, because from the biography it seems that neither Greene nor Sherry had ever been interested in astrology. After discarding 64 trivial events from the 251, no single peak was found, and no peak was observed at the 10:20 am birth time recorded by his mother. Weighting the important events made no difference, nor did changing the rules for selecting hits. Tests with other subjects gave similar results. Worse, charts and events generated at random gave peaks (which are necessarily meaningless) just like real data, see the plots below.

Example plots of Terpstra's test of birth time reconstruction
Left to right: Terpstra's results for Chandu, Greene, and two sets of random data.

Conclusion: birth-time reconstruction based on progressions does not work in general, and consequently those progressions themselves do not work.

A Prediction based on Uranus-Neptune Conjunctions
Sietze van der Tuin AinO 1992, 7.2, 42-45. Uranus is said to indicate change and revolution. Neptune is said to indicate fantasy and idealism. A conjunction occurred in 1820, characterised by technological changes (eg industrial revolution) and political upheavals (eg death of Napoleon), which matches Uranus and Neptune taken together. In 1819 Mars was in opposition to the Uranus-Neptune conjunction, and in the same year panic broke out in financial markets leading to the worst economic depression of the 19th century. We might therefore predict a similar situation for the next conjunction, which occurred in 1990, and for 1992 when Mars is again in opposition. And indeed there was panic and despondency in the money markets in 1992. Critique by Bert Terpstra on page 45: In fact there was only currency speculation in 1992, where speculators exploited EEC rules that force central banks to purchase any EEC currency that is threatened. The speculators made billions in profits, so they were anything but despondent. In any case, currency speculation and depression are not the same.

Saturn-Uranus and the Weather
Marvin V Layman Kosmos 1992, 21.3, 4-8. During 1975-1978 I followed weather records to see what effect the Saturn-Uranus square might have. There seemed to be no correlation, which was surprising because tradition suggests it would consistently bring colder, wetter weather. During 1988 these two planets were conjunct on three occasions, mostly staying within two degrees, but the weather in Oklahoma (where I live) was not consistently cold and wet. In fact the city of Tulsa had the driest year since 1907. Postscript by the editor Susie Porter on pages 9-10: Most astrologers would associate Uranus with drought and wind. [She says nothing about Saturn, traditionally associated with cold and damp.] The Aries Ingress of 1988 for Tulsa would at first glance not suggest drought, but on adding planetary nodes, helio positions, midpoints, asteroids and Arabic Parts the pictures changes. [Keep firing arrows and eventually we will hit the target.] The more we work with astrology the more we realise that its principles show issues but predicting details can be difficult.

Waves and Wars 1700-1992
Robert D.Doolaard AinO 1993, 8.1, 15-22. (An earlier unrefereed version of this article was published in the Astrological Journal of September/October 1993 pages 268-279. It was subsequently refereed by AinO and an improved version appears in this issue along with an editorial comment. - Ed.) During the Second World War the French astrologer Henri Gouchon had the idea of calculating the ten angular separations between the outer five planets (Jupiter through Pluto) on March 21 for each year. He then added up these ten angular separations and plotted the results on a graph, which he called "the cyclic index" or Jupiter wave. For this research I have included the Saturn wave and the Uranus wave, calculated in the same way for Saturn through Pluto and Uranus through Pluto, respectively. My intention is to compare the cyclic index of the last three centuries with the wars occurring during that period. The question is whether more wars, or more serious wars, break out during the downwards phase than during the upwards phase of the wave. Of 22 mega wars, 20 occurred in the downwards phase of the Jupiter wave and only 2 in the upwards phase. A similar but less marked difference was observed for the Saturn and Uranus waves. Editorial comment: In this study no hypothesis has been formulated in advance and hence none has been verified. The obtained results are dominated by 22 mega wars. This number and the number of cycles are too small for a statistical assessment. Nevertheless, the obtained results are remarkable. Only time will tell whether future big wars will begin more often during downward phases than during upward phases. [As noted by the author, the cyclic index and variations have been much studied by others such as Andre Barbault in L'Astrologie Mondiale 1979 and Charles Harvey in Baigent, Campion and Harvey's Mundane Astrology 1984 and 1992. A diagram and further comments appear after the next abstract but one.]

King George III and Samuel Hemmings
Geoffrey Dean Correlation 1994, 13.2, 17-30. The famous story of time twins Samuel Hemmings and King George III made its first appearance in 1822, two years after their death, in a footnote on page 10 of Ashmand's translation of Ptolemy's Tetrabiblos, which was subsequently quoted without acknowledgement in Raphael's A Manual of Astrology 1828, where it reads as follows (Raphael's italics and query):

In the newspapers of February 1820, the death of a Mr. Samuel Hemmings was noticed. It was stated, that he had been an ironmonger and prosperous in trade -- that he was born on the 4th of June 1738, at nearly the same moment as his late majesty George III, and in the same parish of St. Martin's in the fields; -- that he went into business for himself in October 1760, when his late majesty came to the throne; -- that he married on (the 8th of September 1761,) the same day as the king; and finally, after other events of his life had resembled those which happened to the late king, that he died on Saturday, January 29th 1820, on the same day, and nearly at the same hour as his late majesty! QUERY. After such an authenticated and luminous instance as the foregoing, where the lives of two individuals born at the same moment, corresponded in every remarkable particular, even in life and death; can the Astrologer be justly accused of superstition or absurdity, should he pronounce the fates of mankind to be subject to planetary influence? Or can any rational mind, upon mature and sober reflection, attribute the foregoing most pointed agreement in their destinies -- to mere chance?"

A careful search of the main London newspapers for the two weeks following the King's death failed to uncover the alleged notice. However, three of them did contain the following obituary:

Obituary notice of Richard Speer
Obituary from The Morning Advertiser, London, 2 February 1820, page 3 last column.

Ashmand's mention of "in the newspapers [plural] of the month of February 1820" exactly matches the above, but for some reason he and Raphael (and countless subsequent copyists) have the name wrong, and have added things not in the original, namely the birthplace, being prosperous, and the simultaneity of events including going into business. In short, the story seem to be a fabrication. Furthermore, if we consider not just this king and this city but also other notables and other cities, then for the then population of England and Wales (13 million) the probability of such a birth-and-death twin occurring by chance is unremarkable. The answer to Raphael's query is clearly Yes. Follow-up by the same author in Correlation 1995, 14.2, 23-27. There is no mention of Samuel Hemmings or Richard Speer in the baptism and marriage records for London in the International Genealogical Index, or in the parish registers for Westminster (where Samuel Hemmings was supposedly born) or Hammersmith (where Richard Speer lived). They record no marriage on the same day as the King, and the only births on the same day as the King were those of the twins Henry and Thomas Wallington. Other parish registers may solve the mystery, but without knowing where to look, and with no guarantee of success, the task is hardly attractive.

War, prosperity, and the 500-year outer planetary cycle
Rudy Bes Correlation 2001, 20.1, 4-27 and 2004, 22.1, 38-51. Evidence is presented that warfare and economic growth are correlated with an outer planetary wave obtained by adding together the sine of the zodiacal longitude of Jupiter, Saturn, Uranus, Neptune, and Pluto, and then reversing the sign, so eg +2.0 becomes -2.0 [the reason for this reversal is not explained]. During 1500-1950 the correlation with an independent warfare index for 116 global wars was -0.34 (p = 0.01 when corrected for number of tests) for both geocentric and heliocentric waves, and -0.11 (nonsignificant) for 335 international wars. The negative sign indicates that wars are associated with troughs in the planetary wave. Global warfare is predicted beginning in 2032. Individually significant correlations with economic growth indicators since 1750 for Canada, France, Italy, Japan, and the USA averaged 0.22 (not significant when corrected for number of tests). Correlations with Germany or the UK were not individually significant. Graphs of outer planetary fundamental waves from 1450 to 2050 are presented, and general comparisons are drawn between the Renaissance and New Age periods. [Bes's planetary wave predictor and Doolard's cyclic index predictor described two abstracts earlier are shown below for 1650-2050 together with the same predictors minus Jupiter for comparison. The scale for each predictor has been adjusted to make them similar in amplitude:

Cyclic index and planetary wave predictors compared
Behaviour of the cyclic index and planetary wave predictors 1650-2050

The predictors are based on different measures (zodiacal longitudes vs angular separation) used in ways that do not occur in a traditional chart reading. If wars really do relate to downwards phases of the cyclic index and troughs in the planetary wave, the mean value of the planetary wave during downwards phases should be less than the mean value during upwards phases, but the reverse is true for the above plots (0.25 vs 0.10, mean sd 1.63). Which, together with the negative correlation between predictors (-0.21 over 800 data points), indicates that at least one of the predictors is suspect. But all that the authors have observed is a correlation, which does not imply a real link. We know that wars and economic crises occur at erratic intervals, and that combinations of planets can match almost any frequency, especially when the combination and phase are not specified in advance and there is no reason to suspect a link in the first place. So apparently positive results that are actually meaningless are more or less guaranteed. Neither author fits a wave to actual wars to see how it performs compared with planets.]

Astrology and the Sex-Trafficking of Girls in Nepal
Padam Simkhada and Dhruba Simkhada Correlation 2004, 22.1, 52-62. Over 5000 Nepali girls are trafficked for prostitution each year. The business is highly profitable because there is a strong demand for Nepali sex workers in Indian brothels. Many Nepali astrologers believe that some parents will have surrendered their daughter on the basis of astrological indicators. But a sample of 100 Nepali sex workers and trafficked girls showed no links with astrological factors, which therefore cannot be used to identify why some parents give up their daughters into prostitution.

3. Signs (16 abstracts)

Results of Research on Marriage Partners
H.Boning, W van Dam, R.M.M.Hepp, A.Kattenburg, C.Kuypers, and R.H.Smit Tijdschrift Astrologie 1978, 2.2, 2. Harmonious links between the Sun sign of one partner and Moon sign of the other partner are traditionally said to indicate compatibility, and the opposite for inharmonious links. Our research group investigated this claim with a sample of 736 married couples as a check on Kuypers' results with 438 married couples that were given in Rudolf Smit's De Planeten Spreken (Fidessa, Bussum, 1976, page 271). For each angle between tropical signs the total observed counts minus total expected counts (146 or 245) were as follows:

 0    30    60    90   120   150   180    Angle between signs
36 -11 -4 13 03 -23 8 Diff from 146 exp (Kuypers)
-9 0 10 15 -9 -14 9 Diff from 245 exp (our check)

The aspects 0 and 180 occur only once in the circle whereas the other aspects occur twice, so the former counts have been doubled to facilitate comparison. In both cases the results for the square (90) are opposite in direction to tradition. For three angles (especially 0, 60, 120, supposedly the most important) the direction does not replicate. The correlation between observed counts is only 0.15 (p = 0.75). When Sun-Sun and Moon-Moon signs included the outcome was unchanged with a correlation of 0.16 between observed counts (p = 0.73). Astrological tradition was not supported.

Role of Venus and Mars in choice of marriage partners
Cornelis Kuypers Wetenschap & Astrologie 1983, 7.1&2, 16-17. Links between the Venus sign of one partner and Mars sign of the other partner are said to be important for sexual relationships. For each angle between tropical signs in 450 married couples, the total observed counts minus total expected counts (150) were as follows:

  0    30    60    90   120   150   180     Angle between signs
-24 -2 9 -2 3 -5 18 Diff from 150 expected

Counts for 0 and 180 have been doubled to facilitate comparison with the others. For 60, 90, 120 the differences are very small but in agreement with tradition. For 0, 180 the differences are larger but contrary to tradition. There is no clear support for the astrological claim.

Can Self-Attribution Explain Sun-Sign Guessing?
Geoffrey Dean Correlation 1983, 3.2, 22-27. In a test conducted in 1975 by the UK Sunday newspaper News of the World, a panel of four astrologers were able to guess the Sun signs of 8 out of 12 subjects (one for each sign) following one or two interviews lasting no more than five minutes each, vs 1 expected by chance. The test was repeated in 1981 and again the panel got 8 out of 12 right. The subjects were friends of the reporter, who had a great interest in astrology. The panel sat around a table downstairs in the presence of the reporter, and the subjects were brought in one at a time in random order. The subjects not being interviewed sat upstairs and talked. Alcohol flowed freely and the atmosphere, including that of the interviews, was most convivial. The panel could ask what questions they liked but could not mention birthdays. The panel were not told the results until the entire test was finished. On both occasions the panel included Julia Parker and John Naylor, two of the UK's leading sun sign astrologers.

So how did they do it? I was present during the second test and was able to tape the panel questions and talk to the subjects. I found they had an above-average knowledge of Sun signs (for example half had read Linda Goodman's Sun Signs), and several mentioned that it was generally easy to see what the panel questions were getting at. For example: Are you energetic? (Aries). Do you like good food? (Taurus). Are you generous? (Leo). Do you love jewellery? (Libra). Are you ambitious? (Capricorn). Do you have foot problems? (Pisces). The questions necessarily involved topics that feature in Sun sign descriptions, so subjects who knew their own sign details could hardly fail to guide the panel to the right answer. Furthermore, when invited to describe their interests, they could reply at length, when only a small amount of astrological role-playing was needed to provide good clues. For example: "Anything physical attracts me" (Aries). "The most important thing in my life is being a mother" (Cancer). "I like a neat and clean home" (Virgo). "I'd rather not wear a tie" (Sagittarius).

In short, the results seem to rely on self-attribution, the tendency for people who believe in sun signs to shift their self-image in the corresponding direction [see next abstract], which can then be picked up by artful questioning. Indeed, the level of self-attribution, determined by my own questions under less than ideal conditions, successfully predicted the test's outcome for 6 out of 8 hits and 3 out of 4 misses. Sun-sign guessing can be explained by subject selection, self-attribution, and body language.

Self-Attribution as a Moderator Variable in Differential Psychology
Kurt Pawlik and Lothar Buse Correlation 1984, 4.2, 14-30. An English translation by Eve Jackson and Dave Stevens of the original German in Zeitschrift fur Sozialpsychologie 1979, 10, 54-59. A test is presented of the hypothesis that the relationship between astrological birth sign and personality differences in extraversion and neuroticism (as reported by Mayo, White and Eysenck) can be explained in terms of self-attribution of personality. Responses of N=799 adult subjects to two questionnaires (German version of the Eysenck Personality Inventory, and Belief in and Familiarity with Astrology) were analysed in several analyses of variance, with belief in astrology being one of the independent variables. An explanation of the Mayo-White-Eysenck results in terms of attribution theory was essentially verified.

A New Test of the Zodiac Signs
Brian Riley APP 1984, 2.2, 11-14. One of the most frequent objections raised against statistical research into signs is that only single factors are studied and the numerous other factors allegedly influencing a particular trait are ignored. The present experiment was designed to meet this objection by taking account of the sign position of Ascendant, Sun, Moon, Mercury, Venus, and Mars. The polarity and element scores of 24 subjects with extreme scores on the Eysenck Personality Inventory were compared with their E and N scores. The observed correlations were r = 0.02 for E and r = -0.02 for N, which indicate that the traditional claims of links between personality and signs of the zodiac are unfounded.

A Test of Sidereal Sign Delineations
Arthur Blackwell APP 1986, 4.3, 18-21. The Gauquelins applied statistical teats to tropical sign keywords but found nothing that was significant. The author was familiar with the work of two siderealists of renown, Fagaa and Gleadow, so he decided to test their sidereal sign keywords against the Gauquelin trait word list. Sidereal signs are a more ancient idea than tropical signs and are still in use in the East. Nevertheless their results weren't positive either. In the author's opinion the keywords ascribed to the sidereal signs by Fagan and Gleadow are useless and should be abandoned by any siderealist still using them.

A Sun/Moon sign analysis of football notables
Don Woolson Kosmos 1988, 17.4, 15-17. In a previous study of 359 famous athletes I found a significant deficit of Moon in Pisces (16 observed vs 30 expected, p = 0.05), which led me to predict that a similar deficit would be found among professional footballers. But 1210 football players from Ronald Medell's Who's Who in Football failed to confirm it (99 observed vs 101 expected, p = 0.84), and no Sun or Moon sign was significantly high or low, which led me to question my previous findings. Significant findings clearly do not "prove" anything, they can only add support to a particular hypothesis. [In the following issue, a letter from Dr Alan Richter issue complained that no data was presented in the above article as would be required in any scientific journal, to which the editor replied that the author did not submit any data, and in any case Kosmos was not a scientific journal.]

Sun Sign Patterns of 572 Nobel Laureates
Roy Tate APP-NCGR 1989, 7.2, 38-39. I have recently reissued the 4th edition of my book The Astrology of Genius based on 20 years of research. The chi-square test reveals that the odds of this distribution of Sun signs for 572 Nobel laureates is less than 1:500. Francoise Gauquelin's response: Sorry to disappoint you, dear Roy, but your book does not reveal a new astrological law. It describes via Nobel laureates a well-known demographic law affecting anybody in Europe, high-level scientists as well as ordinary people, that results in some months having more births than others. The best way to eliminate demographic artifacts consists in comparing the target sample with a sample of ordinary people born in similar economic conditions, geographical area and period of time. Maybe you will want to do this control without my interference. I would be ready to publish the outcomes of such a control if you take charge of it. [There was no response]

Sun Sign at Birth versus Life Span for 7136 people
Frans Vermeer AinO 1992, 7.1, 10-11. In this study I investigated whether there is a correlation between the Sun sign at birth and the life span of humans. Special attention was given to the astrological postulation that people born under the sign of Capricorn live markedly longer than people born under the other signs. To that end the required data were collected from all tombstones of seven cemeteries in the South of the Netherlands. The dates of birth and death of 7136 people were transcribed. Infants that died within half a year after their birth were exciuded. All these data were then run through a statistical computer program. One-way analysis of variance revealed no correlation between life span and birth sign (p = 0.15). The life span of Capricorns scored below the mean, so the astrological postulation that Capricorns live longest was not supported.

Sign and Branch of Study for 8379 University Students
Jan van Rooij AinO 1992, 7.2, 8-13. In this study the sign positions of the Sun, Mercury, Venus and Mars were compared between 5147 psychology students and 3232 engineering students. According to Jungian theory psychology students should be high on Intuition and Feeling (Fire and Water signs) whereas engineering students should be high on Thinking and Sensing (Air and Earth signs). Expectancies were calculated using birth rates per month for the years 1952-1973 (all students had been born within that period) provided by the Central Bureau of Statistics of the Netherlands. No result was significant. So psychology and engineering students could not be differentiated by their planetary sign positions. Of course any division of frequencies for the twelve signs will show higher and lower values, and astrologers (like most people) tend to assign meaning to such appealing figures, which in this case are at chance level. This may explain the discrepancy between astrologers who insist that astrology works and researchers who consistently find that it does not.

Guessing Sun, Moon, and Ascending signs
Wim Heideman AinO 1992, 7.2, 36-41. Guesses were made of the Sun, Moon, and Ascending sign of a large number of people encountered in everyday life, eg in bars, without any research plan or controls. The guesses were afterwards compared to the actual sign as given by the subject. The guesses were based on the overall likeness between sign and the person's general appearance and behaviour and were not specifically aimed at Sun, Moon or Ascendant. Usually several guesses were made without being told the answer until the end. But sometimes the response was "all wrong, try again", in which case the author kept guessing until he either guessed correctly or gave up. The results were as follows:

                                      Mean guesses   Mean hits
Sign Subjects Guesses Hits per subject per guess
Sun 728 1783 178 2.45 0.100
Asc 254 775 79 3.05 0.102
Moon 81 287 32 3.54 0.111

The number of subjects is smallest for the Moon because people knew their Moon sign less often than the other two signs. Expectancies were estimated by making a random guess by computer for each subject and then repeating it 3000 times. This gave hit rates of 0.0833, 0.0848, and 0.0861, all close to the theoretical expectancy of 1/12 = 0.0833 for a single guess assuming signs are equally probable, and all less than the above observed hit rates. The author briefly considers the effect of self-attribution, unintentional help from the subject's reactions, and recording errors even though he had carefully noted all misses. He concludes that there is a small but significant surplus of hits due to a real capacity to recognise some elements in the horoscope.

[However his expectancies are invalid. (1) No allowance is made for demography (some months have more births than others) and astronomy (some signs have more days than others due to the Earth's elliptical orbit) that can vary Sun sign expectancies by more than 10% (Moon signs are hardly affected), or for differences in ascension time that can vary Ascendant expectancies by more than 50%. In other words Sun and Ascending signs are not equally probable, and inflated hit rates could be obtained simply by substituting the most frequent signs for each guess. (2) If the 12 signs are equally probable and guesses are made without stopping, the probability that one will be a hit is 1/12. But if we stop after N guesses (as in "all wrong, try again"), the probability of a hit among the remaining 12-N signs is not 1/12 but 1/(12-N), which for N = 1, 2, 3, 4 is 9, 20, 33, 50% higher. That the observed hit rates exceed 1/12 is therefore unremarkable. Unfortunately the author ignores these problems, nor did he record the guessing procedure in each case, so his results cannot be evaluated. Nevertheless if the claimed positive result is due to more guesses (therefore more stops and higher expectancies) as opposed to fewer guesses, the observed hit rate should increase as the mean guesses per subject increases, and it does (r = 0.92). Finally a skilled cold reader could pick the correct sign simply by reciting the signs and watching body language. In short, the results do not support the author's conclusion.]

Astrologically Predictable Patterns in Work-Related Injuries
Sara Klein Kosmos 1993, 22.1, 2-4 and 22.3, 21-30. Study was her doctoral dissertation in psychology, University for Humanistic Studies, Del Mar CA, 1992, with 120 references. In 1988 nearly two million people in the USA suffered disabling injuries at work compared with more than 50 million disabling injuries from all causes (car accidents, falls, firearms, poisoning, burns, etc). There is no specific astrological factor related to accidents, only a general one involving hard aspects from transitting planets to natal planets, the generally accepted order of decreasing severity being 90, 180, 0. The only astrological work dealing with accidents is Charles Carter's The Astrology of Accidents (1932), which looked only at natal planets in 168 cases of accident and ignored transitting planets. In my study the experimental data consisted of the birth dates and accident dates of 1023 people in California who had been disabled for at least three months in an accident at work during 1983-1991 and had filed a Workers' Compensation claim. The data was copied at my request from the reports of doctors to whom they had been referred by lawyers presenting their claim. The copyists were told only that it was needed for a statistical study of injury patterns. Cases were excluded if any date was uncertain, if there was more than one injury date, or if the injury was predominantly psychological (eg stress) or had no definite onset (eg lung disease). The sample consisted of 414 English-speaking cases from three independent sources and 609 Spanish-speaking cases (mostly Mexicans) from a Los Angeles clinic that dealt with Hispanics. In both groups the accidents showed a marked tendency to occur when transitting Sun was 0, 90, or 180 to the natal Sun regardless of orb (between groups r = 0.82, df = 10, p = 0.001). If the expectancy is assumed to be N x (total orb)/360, the observed and expected number of such aspects for N=1023 are as follows:

Orb     Obs     Exp    Obs/Exp
15 471 341 1.38
10 330 227 1.45
5 189 114 1.66

By chi-squared test the excess is very highly significant, p being typically <10-8. The excess was about 20% larger for the conjunction than for the rest, showing that the peak time for accidents was around the victim's birthday, suggesting an effect from too many parties or from depression at the thought of getting older. But only astrology seems able to explain why accidents should also peak around 3, 6 and 9 months from the victim's birthday.

[But if accidents are less frequent at weekends, the assumed expectancy may be invalid. Also the birth and accident dates were reported to the doctor presumably some time after the accident and may therefore be less than accurate. Similarly the minimum 3-month period off work required for data to be included is suspiciously close to the period between hard aspects. These points were covered in a follow-up study briefly reported by Zip Dobyns and Mark Pottenger in their house journal Mutable Dilemma 1996, 19.3, 2 and 1997, 20.2, 2, and even more briefly in 1997, 20.4, 2 and 1999, 22.3, 2, in which they tested a large Swedish database of 2865 persons who had suffered critical work-related accidents in 1993. (Further US data could not be obtained due to new privacy laws.) The Swedish data gave "total non-significance" for aspects between natal and transitting Suns on the day of the accident, although the conjunction was again the strongest. Furthermore in Sweden medical needs are met without having to file a claim, and all but 413 of the accidents were timed, indicating careful reporting. "Sara suspects that some of the claims by younger individuals in California may have been fraudulent."]

Jungian Typology and Astrology: an Empirical Test
Jan van Rooij AinO 1993, 8.1, 12-14. We tested the claim that the Sun's element at birth is related to the psychological functions in Carl Jung's typology, specifically whether the Sun in a Fire sign is related to a dominant Intuition function, Earth to Sensing, Air to Thinking, and Water to Feeling. Birth dates and scores on the Myers-Briggs Type Indicator (a personality test that is the standard way to measure the four Jungian functions) were available for 370 subjects (168 males, 202 females). The dominant psychological function showed no significant relation to element (p = 0.34). When the 370 cases were divided into Intuition vs Sensing, or into Thinking vs Feeling, the relation with element was again non-significant (p = 0.36 and p = 0.03 respectively, among three tests the last is not signifiant), the strongest but still very weak relationship being between Feeling and Water. A relationship between Jungian typology and astrological elements was therefore not supported.

A Test of Alice Bailey's Ray Theory of Sun Signs
Peter Niehenke Correlation 1997, 16.1, 29-31. Last year I learned about a study by James David based on Alice Bailey's theory of "Seven Rays". He claimed that, using this theory, photographs of people could be accurately matched to their Sun sign. I had tested similar Sun sign claims before, with negative results, so his claim was hard to believe. I decided to test his claim with about 200 photographs, of which 70% were rejected as too indefinite. The results were almost exactly at chance level. The hit rate showed no tendency to increase with judgement certainty. These results indicate that the theory as applied by James David has no validity. There seems to be no justification for conducting a more elaborate test.

Astrology: is This the Proof?
Geoffrey Dean Correlation 1997, 16.2, 63-66. The above words headlined a two-page article in the UK Daily Mail of 3 February 1998 presenting "the results of an exclusive study that appears to prove your star sign is an accurate indicator of your character". A representative sample of 1092 adults had to say which of 12 unidentified personality profiles (which were actually Sun sign profiles) applied to them. Six signs chose their own profile more frequently than the other signs chose it. The article suggested this was proof of astrology, but it was wrong because although Scorpios (for example) chose their profile more frequently than did the other signs, they also chose 9 other profiles as equal to or ahead of Scorpio. Indeed, only Aquarians put their profile in 1st position, which (1 hit in 12) is what we would expect by chance. But even that solitary hit is dubious because Aquarius was rated first by all signs except Taureans, who rated it second. In the other direction, Gemini was rated last or second-last by ten signs, and even Geminians rated it no higher than sixth. Furthermore, 53% of the women and 26% of the men consulted their horoscope column very or quite often, so many of them were familar with their Sun sign profiles. Therefore the hit rate should increase as the Daily Mail profiles got closer to the traditional profiles, which proved to be broadly the case. So the Daily Mail results confirm what critics have long been saying, and astrologers have long been denying, namely that the reason Sun sign profiles are accepted has nothing to do with astrology and everything to do with social desirability, self-attribution, and other effects. Which did not stop the newspaper's resident astrologer Jonathan Cainer claiming that the study "confirms the relevance of astrology at a time when some folk are a little too keen to dismiss it as an ancient superstition."

The Start of the Age of Aquarius
Nick Campion Correlation 2000, 19.1, 7-16. The notion that the Age of Aquarius is either beginning or is imminent is frequently found in astrological literature. However, there is no agreement on when it begins or how its inauguration is to be calculated. By the normally accepted definition the Age begins when the First Point of Aries in the tropical zodiac precesses either into the equal thirty degree division of the sidereal sign of Aquarius or into the unequal sidereal constellation of Aquarius. The moment this occurs will depend on exactly where the boundary of this sign or constellation is fixed. Opinion varies considerably from school to school and astrologer to astrologer, as witnessed by the wide variations in Ayanamshas (the difference between 0 degrees Aries in the tropical and sidereal zodiacs) in use in India and elsewhere. In addition, western astrologers increasingly point to planetary movements in the tropical zodiac to define the beginning of the Age. Includes a referenced list of nearly 100 dates between 1457 and 3550 that have been proposed (mostly by believers) for the beginning of the Age of Aquarius.

4. Aspects (5 abstracts)

The Validity of Astrological Aspects
Peter Niehenke APP 1984, 2.3, 9-15. As part of the author's PhD thesis, a sample of nearly 3500 persons (obtained by advertisments in a magazine and two newspapers) completed the Freiburg Personality Inventory (12 scales) and a 479-item questionnaire aimed at verifying astrological claims. The result was 3150 usable responses. The questionnaire included items aimed at verifying aspect meanings given in textbooks. For example, in response to the item "I am really not lucky in love", people with Venus-Saturn aspects should tend to say Yes more than those with Venus-Jupiter aspects. It was not possible to investigate all aspect pairs, so my items were limited to aspects between the outer planets Jupiter through Neptune and the personal planets Sun through Mars including Ascendant and MC, and to aspects between Sun, Moon, Venus and Mars. The results were uniformly inconsistent. There were no significant differences in unluckiness in love between people with Venus-Saturn and Venus-Jupiter aspects regardless of orb (1-9 degrees). Nor was there any consistent link with Saturn and Jupiter positions in 5th house (love, sexuality) or 7th house (partners). Contrary to what the textbooks say, people with Saturn aspects did not feel more lonely, unfortunate, skeptical, unhappy, exploited, or cautious than those without Saturn aspects, even when they had up to four Saturn aspects. Nor was there any link with scores on relevant scales such as Depression on the personality inventory. When the sample was divided in half, responses in the first half that seemed to support Saturn tradition did not replicate in the second half more than expected by chance. The results for the other planets were just as negative. Conclusion: the self-descriptions of my subjects do not fit the teachings of astrological textbooks. Even with good Jupiter trines they do not feel more sunny than others. Even with bad Saturn squares they do not feel more depressed than others. As a professional astrologer, I recognise that the negative results are a reality. But the evidence of my success in counselling is also a reality. A world in which astrology exists is surely more enjoyable than one without it. This remains for the moment even true for me!

Astrological Aspects at the Birth of Eminent People
Michel Gauquelin Correlation 1985, 5.1, 25-35. The occurrence of five major aspects (conjunction, opposition, square, trine and sextile) of orb 5 degrees between Sun, Moon, Venus, Mars, Jupiter and Saturn was examined for several groups of eminent professionals totalling 15,334 cases. When all five aspects were combined the mean excess or deficit of aspects involving planets significant in key sectors was 1.4%, less than one-tenth of the mean excess or deficit of their frequencies in key sectors. Overall the occurrence of individually significant results was at chance level. There was no significant correlation (r = 0.19, p = 0.50) with the effect of the same planets in key sectors. Nor do my results agree with the results of an 1977 unpublished study by Dieschbourg.

[Dieschborg looked at multiple-of-30-degree aspects among groups of eminent professionals totalling nearly 12,000 cases with apparently no overlap with Gauquelin's data vs control groups of roughly the same size born in the same period, and seemed to find a weak effect. Thus 6% of comparisons gave significant results for which the mean excess or deficit of aspects between personal planets (Sun-Mars) and outer planets (Jupiter-Pluto) was about 10% of expectancy, albeit inflated by selection. The orb had been adjusted according to aspect to make the overall probability of an aspect = 0.5, and was typically 10 degrees.]

Thus for six mutual groups (military leaders, painters & sculptors, musicians, physicians, scientists) Dieschbourg's mean excess or deficit was about five times larger than mine, but tended to be opposite in direction (r = -0.20, p = 0.70). These results, and the negligible agreement between key sector effects and aspect effects, seems to be precisely what would be expected if aspects are without effect. My results do not seem to support the traditional meanings generally attributed to these aspects by astrological textbooks.

Planetary Aspects: Improved Method Yields Negative Results
Suitbert Ertel Correlation 1988, 8.1, 5-21. An attempt was made to find evidence for relations between traditional aspects and human birth. Gauquelin data on 20,528 eminent individuals representing 11 professions were subjected to time series analysis, segmenting periods of hypothetical aspect influence together with three preceding and three succeeding time periods of equal duration. The frequencies of births within each time segment were superimposed for each profession and for each of 15 aspects (conjunctions, squares and oppositions between Venus, Mars, Jupiter, Saturn). Mean time series obtained for empirical aspect occurrences were compared with mean time series for equivalent random segments as controls. No indication of an aspect influence was found.

Going a Step FARther
Mark Pottenger Kosmos 1989, 18.2, 20-39 and 18.3, 15-30 and 18.4, 15-34. The author describes the development of his FAR (Frequencies for Aspect Research) computer program, which draws graphs showing how the angular separation varies between any two bodies (including ascendant, MC, and nodes) for any time interval. Despite having worked for many years with planet and house formulas, he was unprepared for the extreme non-uniformity (and thus extreme research hazard) that his graphs revealed. Assumptions of uniform motion, even as a first approximation, are on very shaky grounds. Three of his 36 examples are shown below.

Angular separations for Sun-Mars, Venus-Mars, and Saturn-Uranus
Distribution of geocentric angular separations measured daily. Left: Mars conjunct Sun is roughly five times more frequent than Mars opposition Sun. In samples without controls this could be (and has been) mistaken for an astrological effect. Centre: Venus stays close to the Sun, so Venus-Mars is broadly similar to Sun-Mars but with large additional fluctuations. Right: Abrupt changes in frequency will occur over a small range in angular separation if the sample period is short compared to the period of the planetary pair, as here for Saturn-Uranus.

The FAR program provides a visual picture of the non-uniform contacts that can occur between planets and which researchers need to be aware of. It illustrates one of the many ways in which computers have revolutionised the approach to astrological research by eliminating the need for manual calculations. Graphs like the ones above would have been unthinkable in the 1970s. [[Mark Pottenger's virtuoso plots can also be found in the ISAR Anthology Astrological Research Methods 1995. Another example appears in the next abstract]

Sun, Uranus and Demographics
Mark Pottenger APP-NCGR 1989, 7.2, 27-30. The author examines whether the angular separations between the Sun and Uranus could be related to yearly distributions of birth in a way that could bias results. He calculates by computer the distribution of Sun longitudes for the period 1900-1930 when Uranus is retrograde. During this period the only time when Uranus is retrograde is during July-August when the Sun is in Leo, and his plotted numbers confirm that there is indeed a peak around Sun in Leo with a tapering off on both sides. The figure below illustrates the effect for the 1373 Gauquelin professionals born during 1900-1930 with Uranus retrograde.

Sun longitudes with Uranus retrograde during 1900-1930
The distribution shows a peak with Sun in Leo (120-150 degrees)

5. Houses (3 abstracts)

A Test of House Systems
Richard Nolle Kosmos 1987, 16.1, 25-27. The question of the best house system has plagued astrologers for centuries. My study used 375 notable athletes (most of them from the Gauquelin Book of Amnerican Charts) with 468 non-athletes as controls, all with timed births, whose charts were computer-calculated using Placidus, Koch, and Equal house systems. None showed a significant relationship between athletic achievement and house position, although Placidus did best followed very closely by Koch. In contrast, Mars conjunct ascendant or MC with a five-degree orb showed a significant relationship (p = 0.03) consistent with Gauquelin's Mars effect. Further tests are needed, but as they stand the results indicate that house placement is not a significant chart factor.

Testing for 7th House Rulership
Margot Tollefson APP 1995, 11.1, 18-19. Among married couples, do any of the astrological placements of one look like they rule the 7th house of the other? This idea was tested on 2825 couples from the second Gauquelin heredity study made in 1976. A total of 18 placements were tested, eg Sun sign of one = 7th house sign of the other, sign of the Sun's house of one = sign of the ruler of the 7th house of the other. In each case the observed frequency (typically 0.1 per couple) was compared with the frequency in a sample of 2825 randomly matched pairs. None were significant at the p = 0.05 level. Confining the sample to couples with 3+ children made no difference. Answer to opening question: No.

Global Horoscopes
Michael Wackford Correlation 2005, 23.1, 45-64. This paper concludes a 5-part review of horoscopy in the Polar Regions whose aim was to clear away some of the many misconceptions of circumpolar horoscopy, to examine the nature and viability of a number of house systems, and to establish which methods of house division can be successfully applied in the Polar Regions and therefore across the entire planet. Only Equal, Campanus and Placidus deserve consideration. The first presents difficulties when the circumpolar ascendant reverses, the second pretends that all skies are as witnessed at the equator, the last is the only quadrant system that can be applied in polar regions without sacrificing astronomical and astrological integrity. But it can still fail to give twelve unambiguous cusps.

6. Gauquelin effect (11 abstracts)

A Response to Eysenck's Evaluation of the Mars Effect
Prof Marcello Truzzi APP 1984, 2.2, 27. The case for neo-astrological causalities being present in the Gauquelins' work is greatly strengthened' by consideration of the total corpus of their researches. And this context increases the scientific importance and priority their work should be accorded, while also adding to the over-all extraordinariness of their anomalies. On the other hand, the Gauquelin and Eysenck work (though presenting, I think, real and important anomalies) still represents an extraordinary set of claims for which commensurate proof has not yet been obtained. The work is important and should be encouraged. But we need independent replications and the elimination of more "normal" alternative explanations, before neoastrology can gain scientific acceptance. And that is as it should be. True or false, the answer lies in continued investigation and more studies.

Profession and Heredity: Computer Re-analysis and New Investigations
Michel Gauquelin Correlation 1984, 4.1, 8-24. The present study re-analyses by computer all the profession and heredity data published in 12 volumes by the Gauquelin laboratory during 1970-1971. The computer results confirmed the original Gauquelin results, which were based on hand calculations done years ago, but were somewhat lower in significance due to inadvertent bias in the hand calculations when interpolating from tables. Again, only the Moon, Venus, Mars, Jupiter and Saturn were found to have a significant effect, the average excess or deficit in the rising and culminating sectors 1 and 4 for famous professionals being 14% (range 9-28%). Several new analyses are described which lead to a better understanding of planetary effects, as follows. For famous professionals the use of Placidus sectors, and two-hour sectors measured from the rising and culminating points, produced results little different from those using individual rise/set sectors. The area of maximum influence existed almost uniformly throughout sectors 1 and 4 with a fairly marked decrease beyond both boundaries; by comparison the influence of the opposite sectors seems very small. Sectors 1 and 4 were found to be independent and not harmonics of each other. For the heredity data the zone size for maximum effect was found to be that of the plus zones previously used. If a parent was born with one planet in plus zones there was no significant tendency for the child to be born with another planet in or out of plus zones. Overall the study demonstrates that computers are necessary to avoid human bias and to perform investigations far too complex to be undertaken by hand.

How Strong are Planetary Effects for Ordinary People?
Franz Stark APP 1986, 4.3, 12-17 and 1987, 5.2, 15-22, with comments from Suitbert Ertel in 1987, 5.1, 40-42 and an independent replication by Francoise Gauquelin in 1988, 6.1, 19-25. A questionnaire listing five character types was completed by 106 persons obtained via an astrology magazine, friends, and addresses chosen at random from the telephone directory. The respondent had to give their birth data and say if they knew their horoscope, and with another person they each had to tick which types the respondent belonged to. Each type was described by eight Gauquelin planetary trait words. For example Type 2 was "agreeable, compromising, good taste, charming, tries to please, flexible, diplomatic, evasive" (Venus), and Type 3 was "active, impatient, combative, passionate, energetic, direct, self-willed, indefatigable" (Mars). By 2x2 tests (type yes/no vs planet-in-plus-zones yes/no) all correlations (as phi) between type and relevant planet were positive, being generally between 0.3 and 0.4 regardless of knowledge of horoscope and whether the type was indicated by the respondent or another person, and most were significant. Compared with the typical correlation of 0.04 between planet and profession, these correlations are astonishingly high. Comment: In APP 5.1 Suitbert Ertel points out that the results conflict with those of Gauquelin's much larger samples of ordinary people (eg 16,700) which showed no links above chance between character traits and planet, and that self-attribution was not adequately controlled, to which Stark replies that he has started a replication using the same questionnaire but with better controls. Replication: In APP 5.2 Franz Stark presents its results. This time the sample size was 100 persons. The results were less consistent than before (two were negative), and fewer were significant. Nevertheless the Gauquelin effect continues to show up. Replication: In APP 6.1 Francoise Gauquelin repeats Stark's study on a sample of 227 persons (mostly from the USA and Brazil) obtained via publication of Stark's questionnaire in APP and by handing it out during her lectures. When an allowance was made for the number of tests, no result was significant (the highest correlation was 0.2 for Saturn), nor were they always positive. When respondents were allowed to select individual traits as well as whole types, they often ticked only flattering traits and avoided ticking any type, or they ticked all five types. Conclusion: the questionnaire was not good enough. It needed improvement and then much validation with various groups of people. This will be a slow process, but if we want an efficient tool for testing Gauquelin effects, it has to be done. [There were no subsequent studies with an improved questionnaire.]

Planetary Effects Defy Physical Interpretation
Suitbert Ertel Correlation 1989, 9.1, 5-23. The author's previous research with Gauquelin data confirmed the existence of planetary effects for eminent professionals. However, the present research casts doubt on Gauquelin's physical explanation. (1) For sports champions the planetary effect was unrelated to astronomical variables (distance of Mars from Earth, its angular size, apparent magnitude, declination, right ascension, solar elongation, and radius vector). Furthermore the effect did not diminish during Mars-Sun conjunctions. (2) For ordinary people, Gauquelin's claim that geomagnetic activity enhanced the planetary correspondence between children and parents was not supported. Nor did the planetary effect for eminent professionals correlate with geomagnetic activity. It seems that Gauquelin's positive results with geomagnetism are due to random oscillations. (3) Gauquelin's claim that planetary effects decrease after 1950 -- a presumed side-effect of applying obstetric drugs -- could not be verified with professionals' data. However, the number of post-1950 births was insufficient for a definite conclusion. (4) The accuracy of birth-times on official documents increased markedly through decades 1830-1950 but produced no corresponding increase in planetary birth frequencies. In the light of these results, Gauquelin's midwife hypothesis seems to be untenable, in which case an interpretation of planetary effects in terms of physics and physiology must be replaced by something else.

Confirmation of Gauquelin Effects in 1288 Eminent Physicians
Arno Muller APP-NCGR 1989, 7.2, 17-20. Using the official birth-time records of 1288 eminent physicians, the author has tried to replicate the Gauquelin effect. As one of the data sources for the present study had already been used by the Gauquelins themselves, comparisons could be made during the gathering and the data processing for ensuring a correct replication study. The initial Gauquelin data were also rechecked and corrected when necessary, but essentially their degree of significance remained the same. The new Muller data showed results in the expected direction for Mars, but not for Jupiter and Saturn. A subsample of 452 particularly renowned physicians increased the significance of the Mars result. For Saturn, the highest result was observed with French physicians. Therefore overall the present study confirms the Gauquelin observations.

Unresolved Problems (Heredity, Character Traits, Aspects, Signs)
Francoise Gauquelin APP 1992, 8.2, 15-16. Michel's passing away left many problems unresolved. They need adequate scrutiny now if we want to do justice to his remarkable experimental intuition. They are the Heredity Hypothesis (why did a larger sample fail to bring forth the previously obtained results), the Character Traits Hypothesis (why did Michel's and Suitbert's results disagree), and aspect and zodiac sign investigations (new tests by computer-wise astrologers seem positive but erroneous statistics and astronomical and demographic artifacts explain fully the assumed astrological results). [Today, more than 15 years later, these problems are no longer unresolved]

8000 Chart Factors Tested on the Gauquelin Professional Data
Mike O'Neill APP 1994, 10.2, 8-22. The Gauquelin timed birth data for eminent professionals (N=15,942) divided into 11 professional groups (actors, doctors, etc) and 3 combined groups (writers and journalist combined, all data combined, all data combined after weighting according to the hard-evidence-based nature of the profession as revealed by Ertel's cluster analysis) was tested for 7868 astrological factors, or 14 x 7868 = 110,152 tests in all. The factors tested were planets in signs (696 factors), planets in houses (576 factors), aspects (1960 factors), three-planet formations (2478 factors), harmonic aspects (340 factors), midpoints (1716 factors), and planetary qualities (102 factors such as speed, retrogradation, and distance). The statistical significance of each result was determined by a z-test using expected values based on at least four control charts generated within 2.5 years of every tested chart.

Despite the large number of factors only one replicated sufficiently to suggest a testable hypothesis, namely that eminent professionals will have an excess of easy oppositions and kites involving the ascendant and any of the planets. The pilot results suggested that the excess should be about 5% more than expected, but a replication on untested data found a deficit of 2%, so the hypothesis was not confirmed. Interestingly, although the Gauquelin effect in key sectors was included with the relevant factors, the Gauquelin effect did not emerge as strongly as I had expected (although it did emerge), no doubt because the effect is very specific to certain planet and profession combinations that would not have been singled out in my approach. This suggests that the Gauquelin effect may not be more dramatic than anything else, but came about because it was spotted early and then pursued relentlessly. Nevertheless, if some of the other factors involve effects similar in magnitude to the Gauquelin effect, it seems unlikely that they would all escape detection in my tests.

Planets in Semantic Space
Graham Douglas APP 1995, 11.1, 20-31. I believe there is a need for more study of the structure of the Gauquelin results and their implications for astrology. Although M. Gauquelin described inter-relations between the planets Mars, Jupiter and Saturn deduced from trait-word analysis, it never became a central part of his theory. Here I want to make this inter planetary structure [ie the similarities and differences between planetary trait words] the starting point of speculations instead of marginalizing it as at present. [Then follow 17 pages of speculations such as placing the planets in the four quadrants formed by good-bad and hard-soft.] Discussion by Francoise Gauquelin APP 1995, 11.1, 32-33. In previous issues of this journal, I have expressed my regret that Graham Douglas does not go on to prove the various hypotheses he likes to formulate. Without any confrontation with external realities, his idea of "semantic space" and of "inter-planetary structure" risks to remain sterile. Piling up a succession of brilliant ideas on these never verified hypotheses is like building castles in the sand or houses of cards. Did you ever hear about the historical discussion that kept the best theologians in France occupied for months around the abstract problem: To what gender do angels belong ? The final answer to this problem is said to have been: Angels belong to the male gender, the better one of course! But then the unsolvable objection arose: Why do the most respectable religious painters represent angels with the face and the garb of a woman? A quite unanswerable objection! That's what you risk when you rely on theories and never on tangible facts for confirming their adequacy.

Birth Time Precision and the Gauquelin Effect
Suitbert Ertel Correlation 1995, 14.1, 30-37. The Gauquelin effect should increase with increased birth time precision. The data, however, do not bear this out, on the contrary, the effect even tends to diminish with better birth time recordings, at least from AD 1880 onwards. The expected positive effect might have been overridden by psychological-sociological variables depressing the effect and enhancing birth record precision at the same time. Until now, however, such intervening variables remain enigmatic. [But see next abstract]

Attribution: A Pervasive New Artifact in the Gauquelin Data
Geoffrey Dean AinO 2000, 13.1&2, 1-72. The Gauquelin findings are just as puzzling for astrology (eg no link with the Sun) as they are for science (eg no link with physical variables). But they are consistent with artifacts due to attribution (social effects). Thus for Gauquelin's ten professional groups (N=15942), the births on desirable and undesirable days show significant (p<0.01) surpluses and deficits that are generally consistent across professional groups. The mean planetary effect size is larger on desirable days (easily obtained by faking) than on undesirable days (nobody would fake to get one), 0.029 vs 0.012, which should not be if planetary effects are unrelated to social effects. Faked times do not need to be precise, which explains why, contrary to expectation, effect sizes increase with decreasing birth time precision. The 12398 parents and 12550 children from the first Gauquelin heredity experiment show heredity effect sizes for children that are greater for same-sex parents (mean 0.022) than for opposite-sex parents (0.012), which should not be if heredity effects are unrelated to social effects. Social effects also explain why planetary effects seem to disappear in births after 1950. Gauquelin suggested this was because the births had been upset by medical intervention. But the modern demand for official documentation would necessarily prevent faking. The idea that faking is being prevented is more plausible than the idea that all hospitals, all doctors and all midwives are intervening in all births. Nevertheless we cannot conclude that social effects explain Gauquelin's puzzling findings. We can do this only if planetary effects disappear under conditions where social effects are absent, as when parents are excluded from the birth reporting process and the child is ignorant of its birth planets.

Planetary Effects Brought Down to Earth
Suitbert Ertel Correlation 2001, 19.2, 37-46 and 2001, 20.1, 30-41 and 2002, 20.2, 39-48 and 2002, 21.1, 35-39 and 2003, 21.2, 11-21. Dean's requirement that superstitious beliefs be stronger in rural areas is not supported by the results for 7,952 French professionals. On Christian feast days the excess births for 2,390 priests and monks was not significantly more than for 15,942 professionals. Superstition declined steadily from 1800 to 1950 but the avoidance of unlucky days etc did not, nor did planetary effects, therefore avoidance is not a valid measure of belief. A total of 320,817 hospital births in 1987-1994 showed a strong midnight avoidance, disconfirming any link with witches. Faking dates and faking astrology seems more evil and more difficult than faking just one, so planetary effects should be weaker on faked days, not stronger. Adding other planets should increase the correlation with avoidances due to the extra information but it does not. We now know that Dean-type explanations of planetary effects are untenable. Reply by Dean in Correlation 2006, 23.2, 53-57 points out that there is no requirement that superstitious beliefs be stronger in rural areas. The excess births of priests and monks on Christian feast days was 39% and 13%, both in the expected direction. Astrological beliefs were once regarded by the elite as part of science and do not qualify as superstition. The avoidance was only of the witching minute 0:00, not the witching hour, so it was to avoid ambiguity, not witches. To say more faking = less effect does not make sense. Why should adding non-relevant planets provide extra information? Ertel's points ignore crucial results such as my heredity findings, my cluster analyses, and the astonishingly close match to every one of the many Gauquelin puzzles. Counter-reply by Ertel in the same issue pages 58-61 concludes that Dean continues "to propagate views ... despite ... factual contradictions". [For more on this interchange see elsewhere on this website under Gauquelin.]

7. Tests of astrologers (12 abstracts)

Chart Interpretation: An Alternative Strategy for Counseling
Mary Cummings, Melinda Smith, Kristi Lovick, and Paul Crosbie Kosmos 1978, 8.2, 5-26. This study from the University of Montana explores the public acceptance of astrology, and tests the validity and usefulness of chart interpretations. In Part 1 a questionnaire survey of 220 university students showed that 91% had been exposed to astrology (eg through newspapers and books), 64% knew one or more characteristics of tbeir Sun sign, 56% felt that astrology had some validity, and 36% would visit an astrologer if one was available. The students most receptive to astrology were those not majoring in science or business, who agreed that traditional counselling was valid, and whose parents and friends were receptive to astrology. In Part 2 twelve subjects working blind had to pick their own from three chart interpretations based on date, time, and place of birth. The authentic chart interpretation was rated 1st, 2nd, 3rd by 4, 6, 2 subjects vs 4, 4, 4 expected by chance. [In the table of results 6 is given as 7, but the text suggests it should be 6.] The mean rating on a six-point scale (1 Extremely descriptive through 6 Extremely non-descriptive) was 2.46 for authentic charts and 2.56 for the controls. Neither difference was significant although slightly in the right direction. In Part 3 each of the twelve subjects was given a one-hour oral reading of their own chart followed by a questionnaire to determine its usefulness. Nine subjects rated the reading as very accurate, increasing their confidence in the validity of astrology, and rated chart interpretation as very useful in counselling, The other three rated the reading as moderately accurate and chart interpretation as moderately useful in counselling. The researchers noted that the reading was highly effective in eliciting personal and other information from the subject, more than could be obtained in a conventional interview. Conclusion: chart readings are a useful but not necessarily accurate counselling tool. Most important, the counselling relationship rests not on the accuracy of the reading but on the counselling skill of the person doing the reading.

Two Tests of Astrologers
Wout Heukelom Tidschrift Astrologie 1978, 2.2, 12 and 1979, 3.1, 24-25. Four horoscopes (dentist, businessman, painter, engineer) were matched to their owner's profession by 19 members of the NGPA, the Dutch Society of Practising Astrologers. Of the 19 responses, 7 had all 4 correct vs 0.8 expected by chance, 8 had 2 correct vs 4.8 expected, 2 had 1 correct vs 6.3 expected, and only 2 had 0 correct vs 7.1 expected. However, the horoscopes had been provided some years previously by a professional astrologer, and it was suspected that they had been selected because they closely fitted what the average textbook said. For example the businessman had Sun and Mercury in 2nd house, indicating a life dealing with money. So it was decided to repeat the test with the same NGPA members using a new unselected horoscope for each of the four professions. 17 members responded. This time nobody had 4 correct matches, and not many had 2 correct. Indeed, the outcome was so disappointing that the exact numbers were not recorded. The result illustrates the need to avoid bias due to prior selection.

Can Astrology Predict E and N? 2. The Whole Chart
Geoffrey Dean Correlation 1985, 5.2, 2-24. To test whether astrologers using the whole chart can predict E (extraversion) and N (emotionality) in ordinary people, the charts of 160 subjects with extreme scores on the Eysenck Personality Inventory were judged by 45 astrologers from beginners to recognised experts. The subjects had been selected from a parent sample of 1198 subjects to give the 20 most extreme subjects in each of the 8 categories E+, E-, N+, N-, E+N+, E+N-, E-N+, and E-N-, all with birth times generally given to better than half an hour, which allowed a total of 120 E judgements and 120 N judgements, both much larger than the 10 judgements typical of most previous studies. The average pair of opposite extremes was equivalent to the two most extreme persons in a random sample of fifteen adults. This compares with the usual approach in experimental psychology which at best is to take the two most extreme persons in a random sample of three. Allowing 5 minutes per judgement the test required a whole week of evenings to complete and was thus the largest that astrologers were likely to tolerate. All charts were computer calculated by ACS in the style preferred by each astrologer with a choice of several house systems, midpoints, asteroids, and Hindu sidereal with navamsa.

The astrologers judged the direction (high + or low -) of E and N, and indicated how confident they were in each judgement. As a control another 45 astrologers made the same judgements by simply guessing. The result was 5400 judgements each of E and N, and the same number of control judgements. For both E and N the agreement among astrologers was very poor (mean kappa 0.10 for direction and 0.01 for confidence), and the hit rate was at chance level (mean 50.3% vs 51.0% for controls vs 50% expected by chance), showing if anything that judgements were made worse by looking at charts.

Hits out of 120 judgements and variation with confidence
Distribution of hits. Above each plot black circles indicate the equivalent hit rate for judgements made with high, medium and low confidence.

Other things being equal, the hit rate should improve as confidence increases, especially as each astrologer had complete freedom to take into account all relevant factors ranging from birth time uncertainty to uncertainties in interpretation. But judgements made with high confidence were no better than those made with low confidence. Factors such as technique, experience, personality, gender, use of intuition, and birth data accuracy made no difference. The fairness of the test was supported by the high proportion of judgements made with high confidence (34%) and medium confidence (45%). On average each astrologer had 10 years of experience and spent 5 minutes on each judgement.

Palmistry avoids the problem of uncertain birth times, so might palmists perform better than astrologers? This was addressed by a sub-test in which audiences of astrologers or palmists had to make 16 E and 15 N judgements of the charts and hands (both projected as 35 mm slides) of extreme scorers. As in the main test the average time per judgement was about 5 minutes. The 13 most experienced astrologers (all part-time or full-time professionals) averaged 54% hits with an agreement kappa of 0.10. The 14 most experienced palmists (at least 6 were professionals) averaged 51% hits with an agreement kappa of 0.12. The number of judgements was too small to tell if there was a genuine difference, especially as neither group performed better than chance. Synthesis was largely ignored, and audiences were swayed in their judgements by the presence or obsence of relatively few factors such as this aspect or that line. They usually differed on what was relevant, so disagreement was the rule. Thus it was not uncommon for half the audience to vote one way and the other half to vote the other way. This disagreement had no evident effect on their faith in astrology or palmistry.

Do astrologers perform better than cold readers? As it happens an out-of-practice student of cold reading was present at one of the early sub-tests and was in the worst possible position at the back of the room. He noticed that when I announced the judgements to be made, I tended to lower my eyes and voice during the correct judgement (which of course I knew in advance) as if trying to hide it. From such cues alone he scored 72% hits, higher than anyone else even in the main test. After this I took precautions to prevent cues. The mean hit rate was 56% for 17 astrologers before precautions vs 52% for 22 astrologers after precautions, suggesting that astrologers do pick up cues but not to the extent that a cold reader does. The astrologer concerned with maximising client satisfaction could therefore do worse than abandon astrology (but not of course the pretence of astrology) in favour of cold reading.

Summary of Carlson's Double-Blind Test of Astrology
Francoise Gauquelin APP 1986, 4.1, 4-8 followed in the same issue by comments from Hans Eysenck and Teresa Weed Hamilton. The original study by Shawn Carlson appeared in Nature 1985, 318, 419-425 at which time it was the largest study of its kind and was notable for involving an advisory panel of three prominent NCGR astrologers to ensure that the study was fair, and double-blind conditions to avoid any possibility of bias. In view of the controversy it created, the following abstract includes information kindly provided by Carlson in 1986 that is not in the APP abstract or in the original paper. The study involved three different tests as follows:

(1) 128 subjects obtained via advertisements and notices in the San Francisco Bay area had to pick their own chart interpretation out of three. Subjects excluded those who strongly disbelieved in astrology, had previously had a chart constructed, or were under 17. Roughly half were male and the mean age was 28, range 17-65. Subjects could not be test subjects (but could be controls, see below) if their birth place, date and time were not documented and if their birth time was not recorded to better than 15 minutes. The charts were calculated by two astrologers using a Digicomp DR70 [a dedicated chart-calculating computer introduced in the 1970s before PCs became popular], and the interpretations were individually typed by a total of 28 experienced astrologers selected by the advisory panel for competence and a background in psychology. Each interpretation was about 1000 words on pages supplied by Carlson that had pre-printed headings typical of an astrology reading (personality, relationships, career, education, current situation) to ensure uniformity of content and length. So each was representative of the best US professional practice. To avoid give-away clues, each interpretation avoided astrological terms and age indications. In addition, 128 control subjects (same Sun sign as the actual subjects but differing in age by at least 3 years) were given the same task. Usable responses were received from 83 subjects and 94 controls. The 83 subjects ranked the authentic interpretation 28, 33, 22 times in 1st, 2nd, 3rd place, which was not significantly different from the results expected by chance (83/3 = 27.7 times in each case, p = 0.57). The 94 control subjects ranked the authentic interpretation 42, 34, 18 times in 1st, 2nd, 3rd place, which was almost significantly different from the results expected by chance (94/3 = 31.3 times in each case, p = 0.07).

(2) To test whether subjects could recognise themselves, they also had to rank the accuracy of 3 CPI profiles. One was their authentic profile, the other two were chosen at random from other subjects of the same sex. The same 3 profiles were also given to control subjects chosen at random and of the same sex as the test subjects. The subjects had to be matched for sex because the CPI contains scales that discriminate between the sexes. The CPI has 18 scales, 3 of which (well-being, good impression, communality) are designed to detect faking, and the rest provide scores on personality dimensions such as dominance, sociability, self-control, responsibility, achievement, and femininity. The CPI was used in preference to other personality inventories because its scales were judged by the advising astrologers to be closest to what is discernible in a chart. It had also been extensively researched. Unfortunately the subjects had not been advised in advance of this second test, most were not particularly interested in the CPI, and more than half failed to respond. The 56 subjects who responded ranked the authentic profile 25, 16, 15 times in 1st, 2nd, 3rd place vs 56/3 = 18.7 expected, which was in the right direction (and more in the right direction than picking authentic charts) but not significant (p = 0.46). For the 50 control subjects who responded the corresponding rankings were 21, 13, 16 vs 50/3 = 16.7 expected, again nonsignificant (p = 0.61). Given the difficulty of understanding a graph rather than readable text, these results are perhaps unsurprising.

(3) For each of the charts they had interpreted and a further chart, each astrologer was given 3 CPI profiles. One was the authentic profile for the chart subject, the other two were chosen at random from other subjects of the same sex. Each astrologer also received a copy of the CPI interpretation manual that explained the meaning and interpretation of each of the 18 CPI scales. They then rated the fit between each profile and chart on a scale of 1-10. Because all astrologers had some background in psychology (nearly all claimed to have some formal training in psychology, average 3 years, three were professional psychologists, and most claimed to have some experience with the California Psychological Inventory), this test should have been easier than it was for the subjects. Nevertheless, of 226 charts sent out, only 114 were returned. The astrologers matched the authentic profile 40, 46, 28 times in 1st, 2nd, 3rd place vs 114/3 = 38 expected by chance, which was in the right direction but was not significant (p = 0.32).

The advisory panel had predicted that, in tests (1) and (3), the hit rate would be about 50% vs 33.3% expected by chance, whereas the observed hit rates were 33.7% and 34.5%. So even though the panel was satisfied that the tests were fair, the results were at chance level (and still less than the 44.6% hit rate for subjects picking their own CPI profile). Comments by Hans Eysenck and Teresa Weed Hamilton focussed on the CPI. Despite its popularity, most of the validity coefficients for single scales are low, while collectively the scales are both too complex and too limited to be a good test of astrology. Its acceptance by the advising astrologers suggests they had little training in psychology. Carlson's study cannot therefore be considered as a valid test of astrology. [This ignores the first test, which did not involve the CPI, and therefore remains valid.]

Guy de Penguern's New Challenge to CORA
Francoise Gauquelin APP 1987, 5.2, 10-14. In response to Guy de Penguern's claim that he could determine health problems from the birth chart, I submitted to him 100 cases with birth data and approximate birth time, each with the date (sometimes also the time) of death. The cases were taken from the register of a Paris hospital. Initially he gave a dozen health problems for each case, but the number of agreements with the hospital records was poor (6% for the first 50 cases), so he agreed to focus on deaths by cancer and cardio-vascular disease. But for 51 cases he predicted more cancer than expected by chance among non-cancer cases than among cancer cases, which was in the wrong direction. I must thank Guy de Penguern for having had the patience and dedication to thoroughly perform this test. It brings a clarification that had to be done.

An Open Letter to National Public Radio
Purcell Styber Kosmos 1990, 19.4, 34-43. To Morning Edition, National Public Radio, Washington DC. You recently broadcast a tacky segment on astrology that shows you know nothing about it. The astrology you find impossible to accept (newspaper horoscopes) is the same astrology that genuine astrologers refuse to accept. There is a large body of evidence suggesting that genuine astrology cannot be dismissed. Gauquelin's Mars effect, Vernon Clark's matching experiments, John Nelson's radio studies, Frank Brown's work on oysters, to mention only a few. But try it for yourself. Choose two people you know, send me their birth data, and I will send you a four-page psychological analysis without knowing their names or even seeing them. Since people with fixed opinions are rarely receptive to contrary evidence, I will understand if I do not hear from you. [No response was reported in subsequent issues. Of the four examples of evidence, three were known at the time (but evidently not by the author) to rest on artifacts, and the fourth (Gauquelin) requires further work before artifacts can be disregarded.]

The Devil and his Advocate
Wout Heukelom AinO 1991, 6.1, 29-39. The author reviews three large-scale tests of astrologers by Carlson, Dean, and McGrew & McFall. [Abstracts of all three are included in this section.] He plays the role of a devil's advocate who criticises both astrologers and scientists. Main conclusions: (1) Astrologers think they can judge the quality of well designed scientific studies, and carry out difficult matching tests. Apparently they cannot, but neither can any untrained person. (2) In these three studies not a single astrological rule, either alone or in combination, has been found valid, for which there is no satisfactory explanation from an astrological point of view. (3) Readings by astrologers of the whole chart have been found invalid. Another demerit point for astrologers. (4) Astrologers disagree in their assessments of a single birth chart, and in their confidence in that assessment, and more than one third disagree with themselves when assessing the same natal chart for the second time. (5) Astrologers assume their success is due to the validity of astrology. They are wrong, because clients appear just as happy with delineations from wrong charts. (6) Probably astrologers are not aware of the pschychological and physiological mechanisms (like Barnum effects and body language) that could explain their apparent success. (7) Astrologers protect themselves against the negative results of scientific studies in which they took part by adopting astrological and nonscientific explanations.

A Collaborative Vernon Clark Experiment
John McGrew and Richard McFall Correlation 1992, 11.2, 2-10. Six astrologers matched 23 birth charts to comprehensive case files, including photographs and a 7-page 61-item questionnaire devised by them that covered hobbies, interests, school grades, best and worst subjects, talents, vocational interests, past and present jobs, education and occupation of parents, type of neighbourhood where they lived, relationship with parents and siblings, dates of deaths in family, dates of moves across the country, birth defects, disfiguring injuries, deviant behaviour, convictions, whether a victim of serious crime, honesty, main personal problems, religious beliefs, health problems, fondness for travel, phobias, dislikes, attitude to authority, sexual relationships, loyalty, favourite colours, ideal living situation, punctuality, perseverance, a night or day person, strengths and weaknesses of best friend, goals in life, height, weight, hair colour, eye colour, skin colour, and race. Each question was open-ended rather than forced-choice, because the astrologers felt this better represented everyday astrological practice. The charts were of native Americans (4 men, 19 women) aged 30-32 years. Altogether the information provided was considerably more than would be involved in a typical consultation. Birth times were verified by birth records and in most cases were recorded with a precision of five minutes or better. The mean accuracy, i.e. agreement with the case files, expressed as Cohen's kappa (0 = zero accuracy and 1 = perfect accuracy), was 0.02. The mean agreement between astrologers was 0.03 (0 = zero agreement and 1 = perfect agreement). Accuracy was unrelated to confidence or to birth-time precision. In a follow-up study, five groups of 5-6 astrologers each matched a subset of 5 charts, so that collectively all 23 charts were matched. The mean accuracy was -0.15, i.e. worse than chance and in the wrong direction. Although the astrologers had collaborated to make the experiment as fair as possible, the negative findings had no effect on their belief in the validity of astrology.

An Attempt to Predict Accidental Death with Vedic Astrology
John Dudley Correlation 1995, 14.2, 7-11. The predictive qualities of Vedic astrology were tested using 20 pairs of birth data. One of each pair was a real person who had died in a road accident. The other was a fictitious person who acted as a control. In each pair the birth place was the same, and the birth dates were no more than three months apart, as were the death dates. Using Vedic astrology, a form of astrology widely applied in India, the author (working blind) attempted to identify the genuine accidental death. The result was 11 hits and 9 misses, which is not significantly different from the 10 expected by chance.

The Astrotest: A Tough Match for Astrologers
Rob Nanninga, Correlation, 1996, 15.2, 14-20, followed by a comment from Jan van Rooij 21-25. In May 1994 the Dutch daily newspaper Algemeen Dagblad published an article by Martin Boot, a former astrologer, who argued that astrologers cannot predict. In response the astrologer Rene Jelsma claimed "astrologers can really predict". So I decided to resolve this difference of opinion by inviting astrologers to take part in a test. All participants would receive the birth date, time and place of seven anonymous subjects. They would also receive the subjects' responses to a questionnaire devised by the participants. 5000 Dutch guilders (about £2000) was offered to any participant who successfully matched all charts to their owners.

More than 70 astrologers showed interest, sending in an average of ten questions that I synthesised into a list of 25 that covered education, vocation, hobbies, interests, main goals, personality, relationships, health, religion, etc. I also asked for dates of important events, and added 24 multiple choice questions taken from the Berkeley Personality Profile. Eight experienced astrologers checked the result and had no major objections. At their suggestion I added three multiple choice questions covering family background. The seven subjects were born in the Netherlands during 1957-1959 with birth data supported by birth certificates. Subjects with an Ascendant near a cusp (so a few minutes difference in birth time could change its sign) were excluded. As a precaution, the questionnaire and list of birth data were sent to Dutch skeptics who tried to find the matching pairs. Although one scored three hits, there was no reason to suppose that any of the pairs could be identified by using hidden clues.

Of the 44 astrologers who took the test, at least half had read more than one hundred charts and were very experienced, while one-third were frequently paid for their services. One quarter were members of the Dutch Society of Practising Astrologers. Half expected 100% hits, and only six expected less than 60%, so their confidence was high. In fact the best astrologer scored only three hits. Half scored no hits, and the average score was 0.75 hits vs 1.00 expected by chance, giving a mean effect size of -0.04, not even in the right direction. There was no evidence that the most experienced astrologers did any better than beginners. The mean agreement between astrologers was 0.01, almost as if each astrologer had used a random generator for deciding their responses.

I asked the astrologers what factors might be responsible for the disappointing results. They pointed out that the outer planet positions were very similar (nevertheless the charts showed many differences), as were some of the questionnaire replies (for example all subjects claimed to be reliable workers, but again there were many more differences than similarities), and maybe the questions were not always answered truthfully (but why should anyone lie about their hobbies or the date of their wedding?). Some participants felt they did not receive enough information, but nearly all had received more than they had asked for. So these arguments are unconvincing. Even if all responses including birth data were totally false, this would not explain why the astrologers failed to show mutual agreement.

Some leading Dutch astrologers explained the results by resorting to the paranormal. They claimed that astrologers can get hits only by using their intuition or by tuning in to the cosmic order, which can be done only during authentic consultations. As soon as anyone interferes by selecting clients or asking questions, this ability disappears. But clearly the participants would not agree with this view, otherwise they would not have participated. Furthermore their confidence presumably included their confidence in intuition, whose role they would have maximised by their method of working. So the results allow no reason to suppose that astrology depends on paranormal influences. Includes birth data. Comment In this test the astrologers were offered such an enormous amount of information that solving the puzzle was unlikely, yet that was precisely the situation they requested. So their panicky explanations are unnecessary. They simply overrated their ability.

Leo Knegt: A White Crow Beyond our Wildest Dreams?
Rudolf Smit Correlation 1997, 16.1, 3-18 with a follow-up in 1998, 17.2, 72-75. If you wish to upset the law that all crows are black, you must not seek to show that no crows are; it is enough to prove one single crow to be white. William James, What Psychical Research Has Accomplished 1897. Leo Knegt (1882-1957) was one of the Netherland's most eminent astrologers. He was tested in a 1933 blind trial by the lawyer Cornelis Van Rossem, who gave him the birth data and gender (and nothing else) of ten subjects selected for precise birth times, distinct characteristics, and the availability of someone who knew them well. Knegt had to describe characteristics that could be verified, and had to avoid anything that was general, ambiguous, or hard to verify. In some cases Knegt was asked to focus on particular issues such as career, health, or whether the subject had a very unusual character trait. As a precaution, unknown to Knegt, Van Rossem had altered the birth co-ordinates very slightly, which made no significant change to the chart but prevented identification of the registry office and therefore identification of the subjects by inspection of registry office records. In his published results Twee Occulte Problemen (The Hague 1933), Van Rossem reproduces Knegt's ten readings in columns side by side with his own comments and those of the independent assessor.

Knegt's interpretations were found to be both accurate and at times amazingly specific, certainly more specific than most astrologers today would consider possible. For example his correct prediction that one female subject would find a position on a passenger ship could hardly be more specific considering that (1) in those days not many women had paid positions, (2) the subject never had a job before, (3) the position of the stewardess on a passenger ship was certainly much harder to come by than the position of an office clerk, and (4) Knegt did not use Pluto since it had then hardly been introduced into astrology, nor did he use midpoints or other modern techniques, his interpretations being based largely on planets in houses. Similarly he correctly identified a formidable swindler, a prominent inventor, marked nervous disorders, and long-term troubles due to swollen feet. In effect Knegt seems to have been an astrological white crow, living proof of the impossible. So how did Knegt do it?

For its time the test was quite good but by today's standards there are deficiencies. For example, it was not double blind, and it could have been influenced by three kinds of artifact: (1) Those that Van Rossem was unaware of because he was a lawyer and not a psychologist. (2) Those that nobody was aware of because it was the early days of experimental psychology. (3) Those that we ourselves are unaware of because Van Rossem does not give enough detail to answer our questions. For example, was it all done by post, were there time limits, where did he get his data and his second assessor, did anyone check that the printed results were identical to the original records, might Knegt be already familiar with some of the charts (at least four of the subjects were public figures), and so on. We just do not know, so we are unable to draw firm conclusions. In this article I look at Van Rossem's blind trial in detail, giving examples of Knegt's specific interpretations and the charts they were based on, and asking readers for their help in trying to find out how Knegt did it.

I also give the birth data for five subjects to see if readers can match them to (1) the case histories plus (2) correct specifics from Knegt's interpretations plus (3) Van Rossem's comments. For example they have to pick the chart that matched (1) "A public and typical social figure; the native was a well-known speculator; a marked ascent in life followed by an inglorious end." plus (2) "Big spender. Short-lived role as author. Solid relationships in political and economic circles." plus (3) "Everything I asked for was in the interpretation. Even the special talent was hit right on the head. Also the psychological description is excellent." The point is, in each case Knegt was successful and the way he was successful is shown in detail, so astrologers cannot argue that the test is unrealistic. I gave them a year in which to reply.

Follow-up Only two readers sent in suggestions about how Knegt did it. They did not agree on the significators that Knegt might have used, so no conclusion was possible. As for the matching test, I had also advertised it in the AA's newsletter Transit in order to reach as large an audience as possible (now perhaps more than 1500 readers), but again only two astrologers responded. So I appealed to the Australian astrologer Dymock Brose, who had a large worldwide audience with his website and his monthly Astrologer's Forum newsletter, which quickly led to a further 19 astrologers responding. The results were disappointing. Most astrologers scored around chance level, namely one hit only. Only two astrologers reached three hits out of five. Overall there were 29 hits vs 21 expected by chance, whereas if each astrologer had equalled Knegt's performance there would have been 105 hits. Among the 21 astrologers there were 208 agreements, slightly less than the 210 agreements expected by chance. The effect sizes were hits 0.095, agreement -0.002. [In 2006 Smit attempted a further follow-up of the matching test via the internet. Eight astrologers responded. One scored two hits, another scored one hit, the rest scored no hits. Most of them used Chinese astrology.]

Can Astrologers Pick Politicians from Painters?
Suitbert Ertel Correlation 1998, 17.1, 3-8. A German astrologer had published an improved method of chart reading and asked me to test his skill. So I gave him accurately timed birth data for 20 Scottish politicians and 20 Scottish painters, in randomised order, copied from documents at Edinburgh's birth registration office. He was confident of being able to tell which was which, but his judgements were no better than chance. The challenge was then extended by internet communication. Eleven experienced astrologers asked for the birth data and sent in their judgements (five more received the birth data but did not respond). From their comments their methods seemed rather diverse, but all seemed to approach the test in a serious manner. However, they did not perform individually better than chance (range was 7-13 hits, four scored 12 hits, p = 0.25 by binomial test, and one scored 13 hits, p = 0.13). Nor did they succeed as a group (mean 10.7 hits, p = 0.75, equivalent effect size r = 0.068). Moreover, mean agreement was poor (kappa = 0.056, where 0 = no agreement and 1 = perfect agreement). The result is consistent with previous studies. Possible reasons for the failure (other than failure of astrology itself) include wrong birth data, wrong professions, and insufficiently experienced astrologers, but none of these seem plausible. Includes birth data.

8. Approaches to research (16 abstracts)

Unlike the previous abstracts that focus on research results, those below are of articles that focus on approaches to research and related matters such as computers, internet, hidden persuaders, and the need for science. Astrologers see the universe as consisting of connected bits so we can tell what people are doing by looking at the planets. The current scientific model sees the universe as consisting of disconnected bits organised by natural evolutionary processes. Some astrologers claim this difference in viewpoint means that astrology cannot be studied by science. Other astrologers disagree, and this section includes some of their arguments for and against. Be aware that most of their arguments ignore the effect of hidden persuaders and are therefore of historical interest only.

The Whole is More than the Sum of its Parts
Peter Niehenke APP 1983, 1.2, 29-32. Statistical investigations into astrology are often accused of not doing justice to the astrologer's intuitive skill, thus bringing about their bad reputaion among convinced astrologers. But some astrologers have misused their art in a similar way and brought it into disrepute among scientists. Between a justified distrust and the unjustified fear of bad results, we should accept the usefulness of statistical tools for overcoming our prejudices and improving our knowledge. Whatever the outcome of such inquiries, positive or negative, it will help the discussion of what astrology really is. We need to be aware that the multiplicity of horoscope components make it open to any interpretation whatsoever. Two examples: (1) Recently I saw three horoscopes of John Lennon, each with a different birth time, yet each indicating "definitely" Lennon's sudden death. (2) One of my clients had consulted four other astrologers before she came to me. She judged my interpretation to be the most adequate of all, and showed me the work of my colleagues for comparison. I thus became aware that I had made an error of twenty years in her birth date!

Professional Astrology Discussion
Anon Kosmos 1984, 13.3, 28-32. A discussion among nearly 20 leading representatives of several US astrology groups reached agreement on the following definition of astrology: "Astrology is the philosophy that postulates a relationship between relevant celestial phenomena and/or processes and certain terrestril affairs". The definition leaves open the cause of the relationship, and leaves undefined the celestial phenomena and terrestrial affairs to allow its adoption by schools of astrology regardless of the techniques they use. Agreement was also reached on a code of ethics based on the codes of several of the represented groups. [Dictionaries, encyclopedias and astrology textbooks have defined astrology variously as a science, a supposed science, an art, a divinatory art, an art/science, a language, a philosophy, and as a system for self-understanding. But the majority (roughly half) define astrology as the study of relationships between the stars and human affairs. In non-astrology books the reference is usually to supposed relationships.]

Opinions from a German Conference
Francoise Gauquelin APP 1984, 2.3, 3-4. In this changing world astrology has lost its immobility. It has become a research topic for diplomas and theses, more numerous each year, presented at universities all over the world. This research trend is bringing continuously new arguments for and against the traditional laws of the astrological system. How is this new trend perceived by astrologers? An open debate presided by Peter Niehenke at the end of the Conference of the German Astrological Association in Bensheim (13-15 April 1984) asked the question: Is Astrology in a Crisis? The variety of reactions to this question was interesting. Here is my summary of some of them: Peter Niehenke: Is a judgment possible for us? If the result comes out negatively, will we admit it to be true and give up astrology? Herr Schulze: Astrology being a body of knowledge preserved by Babylonian priests on clay tablets, and most of the tablets having not yet been deciphered, how can this body of knowledge be already judged and condemned by modern science? Thomas Koberl: There is never one definitive judgment, but a long series of tests, with the possibility of a slow adaptation of our practice to their verdict. Francoise Gauquelin: I don't consider that as a crisis, only as the normal course of events. When old assertions begin to be verified instead of remaining intangible, some will need modifications. It may be felt as painful, but it is unavoidable. Frau von Kraft: Astrology cannot be in a crisis, because it is not a scientific theory that can be checked. It is just a method for helping others. Gudula Beyse: Astrology being a science of the mind, it cannot be reduced to a physical science. It is based on analogies, not on rational methods. Heidrun Kunzmann: The way in which scientific researchers study astrology is excessively simplistic. This is why it cannot bear results. Thomas Koberl: This becomes evident when astrologers, each using a different method, nevertheless achieve successes in their practice. Karlheinz Grzybienski: Of course we are used to start from the observation of individual cases. But with many individual cases, a general knowledge should ultimately emerge. Adele Conrad: The crisis comes from having so many different methods, so many theories. In this sense, it seems to me that scientific checkings of what is really working may be important. Manfred Groeger: It is important to make things quantifiable, for there is a need for proof. Peter Niehenke: Yes, but aren't we used to do it in an often superstitious way? Have we enough training for making useful comparisons? I have often the impression that we just want to rest on our cosmic bed. This is not enough. We must accept a certain crisis in our too comfortable beliefs.

Personal Computers and the New Astrology
Richard Nolle Kosmos 1987, 16.1, 36-41. Only since the mid-1979s has solid-state technology made it possible for the average person to have a personal computer. Today (this was written in 1986) a PC as powerful as the mainframe computers used by NASA to land a man on the Moon is affordable by any astrologer. An off-the-shelf PC XT takes only 1.5 minutes to calculate and print an error-free natal chart, often with options not readily available in books such as parallax correction. Before computers the Erlewines used punched-card charts sorted with metal rods to find how many had a particular factor, while my wife and I shuffled many hundreds of charts in index-card form to do the same. All of those charts had of course to be calculated by hand. Today a PC XT will take just a day or two to do work that previously took months of painstaking brain-draining exhausting labour. Yet astrologers remain trapped in dogma and tradition. We learned from books written by authors who learned from books, and so on, thus passing on a set of rules accepted more or less on faith, simply because we had no way to test them systematically. It was possible only to apply them here and there using the few charts we could calculate by hand. Today we don't have to accept this mess any more. [PCs in 2007 are a thousand times faster than an XT and can calculate a chart in a fraction of a second]

Methodological Issues
J.E.Becerra Kosmos 1987, 16.2, 42-47. Reprinted in APP 1988, 6.1, 36-39. The author, a medical epidemiologist, evaluates the strength of observed astrological correspondences and thus their relevance to astrological practice. He points out that measures of statistical significance depend on both the sample size and the correlation between variables, so the latter is confounded by the former. What is needed in astrology is a measure of the strength of the association. In epidemiology such a measure is the relativew risk, given by chance of X if you have Y / chance of X if you don't have Y. For example the chance of lung cancer if you smoke is about 10 times the chance if you don't smoke. For the Mars effect the relative risk = the risk of an eminent professional being a sports champion if Mars is in a key sector / the risk of being a sports champion if Mars is not in a key sector. Here the relative risk is given by (Obs/Exp) / (N-Obs)/(N-Exp) where Obs = observed frequency, Exp = expected frequency, and N is the sample size. In terms of the percentages given by Gauquelin, relative risk = 20.8/17.2 / (100-20.8)/(100-17.2) = 1.26. That is, an eminent professional is 1.26 times more likely to be a sports champion if they have Mars in a key sector than if they don't. Since there is no Mars effect for ordinary people (ie Obs = Exp), the relative risk for ordinary people is 1.00. That is, an ordinary person is no more likely to be a sports champion if they have Mars in a key sector than if they don't. For low relative risks, usually defined as less than 1.5, the statistical significance mostly depends on sample size, so even very significant results (as in the Gauquelin work) may have no practical value. The author suggests that a correspondence will be useful in practice only if the relative risk exceeds about 2. Three years later, in Kosmos 1990, 19.1, 2-4, the author notes that nobody has heeded his advice, adding that the improper use of statistics by some authors is not a valid argument for ignoring scientific standards in astrological research. [Unfortunately relative risk has no upper numerical limit, a disadvantage that can be avoided by using instead the effect size expressed as a correlation, which has limits of ±1. For the Mars effect it is given by (Obs-Exp)/(N-Exp) = 0.043 or very roughly by (relative risk - 1) x Exp/N = 0.045.]

Research: Some Concepts
William E Brandt Kosmos 1988, 17.4, 4-12. The term "research" has been applied so widely that it now has no single identifiable meaning. One approach classifies it as historical (looks at past events), descriptive (looks at present events), and experimental (tests ideas). Regardless of the type, good research is careful and systematic. It cannot tolerate anything vague, sloppy, or unclear. Research is needed if astrology is to attain credibility. But first the aim must be defined. Research is not the answer if we fail to define the question.

Science as a Way to Consensus
Wout Heukelom AinO 1990, 5.2, 1 & 6. In the spring of 1987 a successful "Studium Generale" was held on astrology at the University of Groningen. Yet beforehand there had been complaints from both sides. Scientists objected (wrongly) that astrology was unverifiable. Astrologers felt (wrongly) that speakers would be stacked against astrology (in fact just as many astrologers as scientists were invited to speak). The organisers wished only to have a fair discussion about the merits of astrology based on the facts gathered through research. This is the only way to a fair consensus. We should not be for or against astrology, we should want only to get the facts right.

Unprofessional Tendencies in Astrology
Jacob Ruijling AinO 1992, 7.2, 26-35. Why do outsiders perceive astrology as nonsense? For example the Dutch skeptics society want a disclaimer in the horoscope columns of newspapers. One reason for this perception is that astrologers ignore the results of empirical research, yet anyone who studies this research must conclude that astrology cannot live up to its claims. So there is a tension between astrology and observation. For example a joint committee of Dutch astrological organisations, set up to formulate guidelines for improving astrological practice, concluded in 1992 that the quality of a birth chart reading comes from the astrologer and is not determined by any particular technique. This immediately creates problems because ethical codes imply that astrologers can predict the future (eg of the client's health) whereas the research shows that astrology cannot predict. The same codes say that astrologers should not predict, but by definition astrologers cannot avoid predicting if they use birth charts. Important questions that I cannot answer include whether education (eg in medical astrology) can be tested if the discipline itself has not been tested, how education, diplomas, and discipline are connected, and what kinds of education presently exist. If astrologers are to progress they must recognise that (1) astrology cannot predict, (2) astrologers predict but not on the basis of astrology, and (3) astrologers should focus more on what makes astrology appear to work and on how astrologers should work.

Difficulties in Attempting to Study the Meaning of Pluto
Marian van Brakel AinO 1993, 8.1, 36-41. This article illustrates the difficulties one can encounter when attempting to do astrological research. In my astrological practice I had become intrigued by the link that transitting Pluto seems to have with difficult periods in the lives of my clients. So in March 1991 I joined a committee to study the effect of Pluto. We advertised for subjects willing to be interviewed and collected 13 useful charts. But not everyone could interview them, we were inexperienced in research, we could not agree on the results or on how to proceed, astrologers who knew about science criticised our approach, and ordinary astrologers wondered what we were doing and thought it was all a waste of our time. The committee was dissolved in April 1992. The article ends with a plea for suitably trained people to start a new research project that avoids the above pitfalls. [But they already had, see first section]

Is the Scientific Approach Relevant to Astrology? (Key Topic 1)
Geoffrey Dean, Arthur Mather, and 33 others Correlation 1994, 13.1, 11-52. To collect views on this topic we (Dean and Mather) compiled a brief discourse that explored the scientific approach, its limitations, and its relevance to astrology, and sent it to over 100 potential commentators (mostly astrologers, mostly non-subscribers to Correlation, and about 50% female), of whom 23 responded. Their opinions ranged from "excellent and provocative" to "virtually valueless." We revised the discourse in the light of their comments and sent it back to the 23 respondents (13 replied). We also sent it to several key astrologers who had not responded to the initial discourse (2 replied), and 10 interested scientists (8 replied). Where possible each reply was recycled with its author to eliminate avoidable arguments and to cover questions that informed readers might be expected to ask. In total 33 people responded of which 4 were female. Their contributions ranged from a few sentences to nearly 4500 words, each bringing a view that, collectively, provided a breadth never before achieved in print. From now on nobody should debate scientific issues in astrology without first reading this collection. Summary: (1) Is the scientific approach relevant to astrology? Yes, but only to those parts testable by observation. No assumptions about how astrology works (eg causally or non-causally) are required. Thus to test whether a person fits his chart better than a control requires no assumptions whatever. (2) Why are scientists and astrologers in conflict over whether astrology works? Mainly because they tend to look at different things. Scientists are mostly concerned with accuracy (controlled tests) whereas astrologers are mostly concerned with satisfaction (client acceptance). But accuracy is unrelated to satisfaction. So their views can conflict yet both can be right. In particular cases a more important reason on either side may be dishonesty, ignorance and arrogance.

Different Approaches to Astrological Research
Charles Harvey Correlation 1994, 13.2, 55-59. The ordinary research done by astrologers this century has led to conceptual breakthroughs such as midpoints, harmonics, composite charts, and Astro*Carto*Graphy, whereas the formal research by researchers has led to something of an impasse. Why the discrepancy? Currently the only approach seen as having scientific value focusses on measurement (Earth) and analysis (Air). Other approaches such as intuition (Fire) and feeling (Water) are generally ignored, yet these qualities underpin astrology. So the discrepancy arises from the failure of researchers to compare like with like. For example it might be fruitful to compare chart readings not with personality scores but with the profile produced by a depth psychologist or a psychologically aware journalist. Or to have a psychologist evaluate the reading of someone she knows intimately. Such approaches may not lend themselves to number crunching but they do involve comparing like with like, and therefore have a greater chance of revealing astrology's power to describe people. The more we think in the symbolic terms used by astrology the more likely we are to come up with convincing results.

Authenticity a Precondition for Tests of Astrology
Leon van Assem, Wil Rozenbroek and Steef van der Weele AinO 1995, 10.1, 6-8. In 1992 NGPA, the Dutch Society of Practising Astrologers, surveyed the philosophical opinions of its leading members and received 19 replies. Concerning scientific research into astrology, the views they agreed on were as follows: The birth chart shows a person's inner life, disposition, health, present circumstances and future tendencies, but not abilities, morals or IQ. Not every chart indication will unfold. Indications can be overruled by free will. There is an ordering principle in life (as above so below) that is thought to be working during an astrological consultation. A successful consultation requires intuition, rapport, and conversation with the client to identify problem areas in more detail and to correct errors in the astrology. The last means that contradictory indications by different systems do not matter. Only a few respondents felt that chart interpretation without knowing the client was possible and useful.

Some Philosophical Problems of Astrology (Key Topic 2)
Geoffrey Dean, Peter Loptson, and 13 others Correlation 1995, 14.2, 32-44. Modern philosophers generally accept astrology as a source of sympathy and support, but they reject it as a source of knowledge. This matches the idea that astrology can be viewed in two ways, one in terms of the satisfaction enjoyed by users, and the other in terms of its accuracy. Of course the two viewpoints may not be independent but for the present purpose this is of no consequence. Astrology from the satisfaction viewpoint is generally unproblematic: (1) Satisfaction typically rests on value judgements and subjective feelings, both of which can legitimately differ. So arguments about the extent and type of satisfaction provided by astrology may be pointless. (2) The astrology so viewed need not be true and is therefore uncontroversial. (3) Nevertheless problems can arise if astrologers needlessly embrace assailable arguments. Why undermine uncontroversial claims with assailable arguments? (4) Problems can also arise if satisfaction depends on perceptions that are in fact false. Action based on false perceptions could be harmful. Astrology from the accuracy viewpoint faces numerous problems: (1) Astrology is defined as precisely not the result of any means we know of. (2) Astrological effects are essentially statistical, are nonidentifiable except after the event, and therefore cannot be an independent source of knowledge. (3) Astrologers have been reluctant to describe what their model predicts, the criteria by which it could be tested, and the evidence they would accept as showing it had failed. (4) No claims to accuracy can be justified unless astrologers make proper experiments and distinguish between alternative explanations and have independent reasons for thinking that astrological effects exist.

Astrology on the Internet
Joanna M Ashmun Correlation 1996, 15.2, 35-51. The author surveyed astrology on the Internet during 1996. Almost all public astrology on the Internet is social or commercial talk of little interest to researchers. Hundreds of serious astrologers subscribe to mailing lists, some moderated and some not. The most striking quality of lists is that they discuss everything, often generating more heat than light, but saying a lot about the power of astrology to generate mutual interest across social and cultural boundaries. As people join and lose interest (it usually takes about 3 months), the newcomers bring up the same old topics, so the level of discussion does not deepen over time. Nothing is resolved or changed. The way astrologers treat researchers and skeptics is just the way they treat other astrologers who disagree with them -- they continue on as if the disagreements never existed. From experience with lists on other topoics, I have to say that posts by astrologers are less literate than average. Most would rather have an iffy quotation from Rudhyar or Jung to support their opinions than some good numbers. Many astrologers have university educations, quite a few have graduate degrees, and they must, therefore, have had to do some reading, writing, and scholarship sometime, but these presumed skills rarely escape to the mailing lists.

I've recommended Correlation a few times when relevant but haven't seen it mentioned except in replies to me. Overall, journals of all types incuding newspapers and popular magazines are mentioned on the lists no more often than once in every 2000 postings. There are frequent mentions of a big variety of astrology books but no agreement that any particular author or book is the last word on any topic. In June 1995, before my present survey, a poll was taken on alt.astrology asking which five astrology books readers would pick if five were all they could keep. The report doesn't say how many replied, but the results require at least 25 respondents if everyone gave five titles. Of 107 titles named, the two most popular books (number of picks not reported, but at least three each) were Reinhold Ebertin's Combination of Stellar Influences and Robert Hand's Horoscope Symbols. Twelve other books were named more than once. In A-Z author order they were John Addey's Harmonics in Astrology, Stephen Arroyo's Astrology. Karma and Transformation, Geoffrey Dean and Arthur Mather's Recent Advances in Natal Astrology, Liz Greene and Howard Sasportas's The Development of Personality Seminars in Psychological Astrology I, Liz Greene's The Astrology of Fate, Liz Greene's Saturn. A New Look at an Old Devil, Robert Hand's Planets in Transit, Marc Edmund Jones's The Sabian Symbols, Debbi Kempton Smith's Secrets from a Stargazer's Notebook, Tracy Marks's The Astrology of Self-Discovery, Dane Rudhyar's The Astrology of Personality, Bil Tiemey's Dynamics of Aspect Analysis. Another 93 titles were named once each.

Astrology and Human Judgement (Key Topic 4)
Geoffrey Dean, Ivan Kelly, Arthur Mather, and 5 others Correlation 1998, 17.2, 24-71. Unless we understand the judgement processes that underly chart interpretation and its assessment, we cannot hope to understand astrology. In the psychological literature these processes have been studied for more than a quarter of a century, but in the astrological literature they have been almost completely neglected. The bad news is that, from start to finish, astrology involves the kind of judgements that we are not very good at. There are many non-astrological reasons [described elsewhere on this website as reasoning errors or hidden persuaders, see Index] why astrology should be seen as valid, none of which require that astrology be true, which of course is not a problem peculiar to astrology. The good news is that there are ways of avoiding the known ways of fooling ourselves, and ways of dealing with crooked arguments, all described at some length. However, such matters are almost universally ignored by astrologers and their teaching institutions. Until the situation improves, the education of astrologers will be fatally deficient.

On Being Properly Scientific
Geoffrey Dean Correlation 2003, 21.2, 43-45. Everyone might agree with the editor in 21.1 that well-designed studies should do away with erroneous conclusions. But it is hard to agree that current studies in Correlation reach this standard. Their conclusions may not be erroneous in themselves but they mislead when they assume that confirmation of astrology's reality is only a matter of time. Indeed, Correlation's current sub-title (research in astrology) would seem to require it, whereas the sub-title before 18(1) (research into astrology) did not, in the same way that we can do research into phlogiston but not in it. The crunch question: Can astrologers do what they claim? This was a difficult question to answer in the 1970s when only a handful of studies existed, but today there are something like a hundred studies (albeit not all of them easily retrievable) with enough variety in design to satisfy most enquiries. Provisional bounds can now be put on what astrologers can and cannot do. The problems with the way astrologers draw conclusions from experience are now known. We can now decide which topic will be of most value to that forgotten person in research, the working astrologer, be it techniques or counselling skills or (worst case) how to survive disconfirmation. Yet these important advances are not mentioned in current issues of Correlation. Authors seem not to be doing their homework. If we are to be properly relevant we must start from what astrologers can actually do under artifact-free conditions, not from what they say they can do under who-knows-what conditions. We must focus on what astrologers would accept as disconfirming evidence, not on evidence they would reject as irrelevant. We must distinguish between reality and vested interests (think of Galileo). Finally we must do what few authors and referees seem capable of doing, namely attend to all the evidence, not just the parts selected to prove a case. In short, we must be properly scientific. Rabbit stew will not be possible if all we have are reports of rabbits. As Mrs Beeton might say, first catch your rabbit.

From       110m 9g 210kb       Home       Fast-Find Index