The claim is based on some of the statistical analyses of Michel Gauquelin, who devoted much of his life to trying to determine whether astrology has any scientific validity. Of all the many analyses performed by Gauquelin, all failed to conclude that astrology had any such validity - with the exception of the one statistical analysis that forms the basis of what became known as the Mars effect. In this one case, the underlying results are anomalous. Because of this, the Mars effect has been the subject of considerable controversy. However, it is important to note that the Mars effect is unknown to traditional astrology; prior to the controversy, no astrologer had ever used or even suspected the existence of such an effect - indeed, Mars in the given position should not relate to sports achievement at all.
A critical paper on this subject was published in an issue of Skeptical Inquirer, the journal of the Committee on Scientific Investigation of Claims of the Paranormal. A Dennis Rawlins responded to this article by claiming that the authors deliberately misrepresenting the purpose of the Zelen test. However, Rawlins' criticism was based on a misunderstanding of the CSICOP article. Those who study the articles on Gauquelin and the Mars effect may note that misunderstandings and miscommunications on almost everyone's part were the order of the day. The misunderstandings led to some heated arguments, fights, broken friendships, published attacks, and tarnished reputations -- all of which, in retrospect, could have been avoided.
| Table of contents |
|
2 Related Topics 3 Further Reference 4 External links |
The Belgian Para Committee, which had tested the Mars effect in the 1960s, suspected that the slightly skewed percentages may have been an artifact.
To test this suspicion and eliminate for bias, Professor Marvin Zelen proposed that Gauquelin randomly pick 100 athletes from his group of 2,088 and check the birth/planet correlations of all the other babies born at the same times and places. This was NOT a test of the Mars effect, but a test of the base-rate (chance) expectation. (The 100 random athletes later evolved into a subsample of 303 athletes.)
Mr. Rawlins thought the above test not worth doing, for as he stated: " . . . we find an inverse correlation between size and deviation in the Mars-athletes subsamples (that is, the smaller the subsample, the larger the success) -- which is what one would expect if bias had infected the blocking off of the sizes of the subsamples" (The Zetetic 2, no. 1, Fall/Winter 1977, p. 81). He also thought, correctly, that general readers would mistakenly think the Zelen test was a test of the Mars effect, not of the base rate.
Mr. and Mrs. Gauquelin performed the test that Professor Zelen proposed, and concluded that the chance Mars-in-key-sector expectation for the general population (i.e., nonchampions) was indeed about 17%, rather than the 22% observed for athletic champions.
In his rebuttal to the Gauquelins' published conclusion, Marvin Zelen analyzed the composition, not of the resulting 17,000 nonchampions, but of the 303 champions. He split the subsample by eliminating the women and by dividing the remaining athletes into city/rural sections and Parisian/non-Parisian sections. (His rebuttal was also signed by Paul Kurtz and George Abell.)
It was this sample-splitting that so shocked Dennis Rawlins and numerous other critics. It appeared that Zelen ignored the purpose of the test, which was to check the base rate of 17,000 regular folks. It appeared also that he was trying minimize the significance of the Mars/key-sector correlations with athletes by invalidly splitting the subsample of 303. It appeared that he was trying to make it look as though the Zelen test was not a test of the expected base rate, but a test of the Mars effect.
Zelen's (and Kurtz's and Abell's) rebuttal was poorly written. Read it a second, third, or fourth time, though, and you will see that the intial appearances are deceptive. Zelen split the sample not to examine the Mars effect, but primarily to examine the randomness of the subsample of 303 champions. And it turned out that the Gauquelins did not choose their subsample randomly. As they admitted, they had trouble finding sufficient same-week and same-village births to compare with champions born in rural areas, so they chose only champions born in larger cities. (If the 22% correlation was an artifact partly based on, say, rural recordkeeping, this would be blurred in such a nonrandom selection.)
Further, the Gauquelins' original total list of about 2,088 champions included exactly 42 Parisians. Their subsample of 303 athletes included exactly 42 Parisians. Now, Paris is divided into 20 sections, called arrondissements, and different economic classes and different ethnic groups inhabit different arrondissements. The Gauquelins compared the 42 Parisian champions (who had been born throughout Paris) to nonchampions of only one arrondissement. (If the 22% correlation was an artifact partly based on, say, economic, class, or ethnic differences in birth patterns, this would be blurred in such a nonrandom selection.)
A history of the affair
Related Topics
Further Reference
External links