Terence Hines - Pseudoscience and the Paranormal

The results of both parts of the study provide no support for astrology. Subjects in the first part of the study were unable to pick the interpretation of their horoscope from the interpretation of two other individuals’ horoscopes at a rate above chance. In the second part of the study, the astrologers were not able to match the horoscope of an individual to his or her CPI personality profile at a rate higher than chance. Given two opportunities to provide impressive empirical support for the reality of astrological claims, in a test that respected astrologers agreed beforehand was fair, astrology failed.

Astrologers claim that numerous other personality traits—in addition to physical characteristics, occupation, and medical disorders—are influenced by one’s sun sign. Culver and Ianna (1984) have summarized several studies that examined these predictions. The predictions have been shown to be wrong. For example, of sixty different occupations studied, not one showed any influence of sun signs. That is, members of these sixty occupations were no more or less likely to be born under one sun sign than another. Among the occupations examined were those of actor, pilot, artist, astronomer, banker, baseball player, chemist, teacher, journalist, lawyer, doctor, opera singer, poet, politician, psychologist, and priest. Among the physical characteristics not related to sun sign were blood type, baldness, hair color, height, sex, handedness, and weight. Medical disorders found to be totally unrelated to sun sign included acne, allergies, diabetes, Down’s syndrome, heart attack, infant death, leukemia, lung cancer, multiple sclerosis, stillbirths, stroke, and muscular dystrophy. Age at death is also not related to sun sign. Finally, the following personality traits are not related to sun sign: aggression, ambition, creativity, feelings of inferiority, integrity, intelligence, leadership, self-expression, sociability, tough-mindedness, understanding, and wisdom. All these findings are in strong contrast to the empirically unsubstantiated claims of astrologers. Those claims can now be seen to be simply wrong.

There is more to astrology than sun signs, and the process of calculating a complete and accurate horoscope certainly takes considerable mathematical ability and training. Numerous variables in addition to the sun sign are said to influence the individual. Among these are planetary conjunctions, houses, ascendants, the relationship between the planets, and the planets’ positions in the various signs. Culver and Ianna (1984) have calculated that there are some 10 35possible astrological predictions. This huge number compares with only 10 27grains of sand on the earth. Some of the predictions involving factors more complicated than sun signs have been tested (Dean 1977; Culver and Ianna 1984; Dean 1986–87a, 1986–87b) but, again, these tests provide no support for astrological claims.

When confronted with the negative outcomes of numerous studies, astrologers sometimes argue that one cannot reject the theory until all, or at least a majority, of its predictions have been tested. This is an unsatisfactory defense of astrology for two reasons. First, because there is such a vast number of possible predictions, it would take forever to test even a small percentage of them. Thus, astrologers will always be able to claim that the theory has not been shown to be incorrect even if every future test of astrological predictions shows the predictions to be wrong. Second, as we have seen, it is unnecessary to test every prediction of a theory to show the theory to be incorrect. If the theory’s major predictions are incorrect, the theory is rightly rejected by scientists and time is not spent testing endless minor-detailed predictions. In the case of astrology, the theory’s major prediction is that sun signs influence personality, occupational choice, and so forth. These predictions have repeatedly been shown to be dead wrong.

The basic idea behind the lunar effect—or “moon madness” or the “Transylvania Hypothesis” as some more lyrical writers term it—is that the moon in its different phases exerts a strong influence on human behavior. It is especially held that the full moon accentuates or increases the probability of all sorts of odd and troublesome behavior. Suicides, admissions to mental hospitals, arrests for public drunkenness, and crimes of various sorts are all said to increase when the moon is full (Lieber 1978). It is also widely believed, especially among maternity ward personnel, that more babies are born when the moon is full than during the other phases of the moon (Abell and Greenspan 1979). The moon’s gravitational influence is usually the mechanism used to explain the alleged effects of the full moon. After all, proponents say, the moon’s gravity influences the oceans, which are largely water. Therefore, since the human body contains a great deal of water, the moon’s gravity must also influence the human body. This, in some unspecified way, results in moon madness. But in fact the moon’s gravitational influence on the human body is infinitesimal—equivalent to the weight of a single mosquito being added to the weight of a normal individual. Gravity is a weak force. As you hold this book, you are outpulling the entire planet Earth.

Campbell and Beets (1978), Abell (1981b), and Rotton and Kelly (1985) have reviewed the considerable number of studies that attempt to link the phases of the moon, especially the full moon, to human behavior. All three reviews conclude unequivocally that the phase of the moon does not influence human behavior. For example, Abell and Greenspan (1979) studied all the births that took place at the University of California Hospital, Los Angeles, from March 17, 1974, through April 30, 1978. During this period there were 11,691 live births, of which 8,142 were natural in that neither drugs nor caesarean section were used. There were 141 multiple births among the live births and 168 stillbirths. Analysis of this huge number of births showed no effect of phase of the moon in any of the four groups of births (all live births, natural births, multiple births, and stillbirths).

Studies of other variables have also failed to find any effect of the full moon. Rotton and Kelly’s (1985) paper reviews numerous published studies that showed the full moon did not influence (1) homicide rate; (2) other criminal offenses; (3) suicides; (4) psychiatric disturbances; or (5) psychiatric admissions to hospital.

Occasionally, of course, a study will report some sort of relation between the full moon and some variable. These studies require close examination. Under such examination, methodological or statistical flaws have appeared that invalidate the conclusions. Templer, Veleber, and Brooner (1982) found that highway accidents at night were more frequent when the moon was full. But their data showed no effect of the phase of the moon on daytime accidents. They devised a rather fanciful explanation of these results based on the effect of moonlight on the human pineal gland. However, as Rotton and Kelly (1985) point out, “a disproportionate number of full-moon nights fell on weekends” during the period studied by Templer, Veleber, and Brooner (p. 292). Templer, Brooner, and Corgiat (1983) reanalyzed their data and, this time, took into account such variables as weekends, holidays, and such. They found that the supposed effect of the full moon on accidents disappeared. Thus, their original finding was due to the effect of weekends on accident rates, not to any effect of the full moon.

Arnold Lieber (1978) has been a strong proponent of the reality of lunar influence on human behavior. He claims to have found a relationship between homicide rates and the full moon in Dade County, Florida, and Cuyahoga County, Ohio, with more homicides taking place when the moon is full. However, Rotton and Kelly (1985) point out a fatal flaw in this data. In attempting to find an effect of the full moon, Lieber and Sherin (1972) conducted ninety-six different statistical tests on their data. They tested the effect of the moon on homicide rates by looking at these rates for, among others, “the three days before and after, the three days before, the three days after, two days before and after, two days before, two days after, one day before and after, one day before, one to two days after, and one to three days after full moons” (Rotton and Kelly 1985, p. 293). Of the ninety-six analyses, three reached the accepted “.05” level of significance, meaning that such a result would be expected by chance only 1 time in 20, or 5 times out of 100. But, if one conducts ninety-six statistical tests, one would expect that 4.8 of them (96 x .05) would reach the .05 level by chance alone. Lieber and Sherin’s (1978) data would provide evidence for the reality of lunar effects on homicide rates only if about ten of their ninety-six different tests showed a significant result. Sanduleak (1984–85) analyzed all 3,370 homicides that took place in Cuyahoga County from 1971 through 1981. He found no lunar influences.