Review of “Redheads and Mars”
Through a series of experiments conducted in the 1980s and 1990s, Judith A. Hill and Jacalyn Thompson investigated the astrological relationship between people with naturally red hair (“redheads”) and the position of Mars in natal charts. This research is particularly interesting because it associates implicit astrological properties with an explicit physical genetic trait. This review and discussion is based mainly on the content of Hill’s 1996 article “Mars and Redheads” that summarized the research program.
Hill and Thompson reviewed the medical literature on redheads and found that red hair has been linked to tendencies toward hyperactivity (excitability, inappropriate aggression, disdain for affectionate conduct, and abnormal activity) and a tendency to suffer from extended bleeding. The astrological literature provides various sources (Ptolemy, Sepharial, Cornell, and Addey) who claimed a relationship between Mars and these hyperactive characteristics as well as other potential expressions, specifically including red hair. Hair color is a feature of personal appearance and astrology traditionally links appearance to the rising point (Ascendant) in a nativity. Additional astrological factors contributing to red hair have been suggested by various authors but Mars rising is fairly consistent in the literature. These sources and tradition led Hill and Thompson to formulate their research hypothesis which is that people with red hair are born more frequently with Mars within 30 degrees of rising and less frequently with Mars within 30 degrees of setting as compared to a randomized or non-redhead population (Hill; 1996, 31).
Hill and Thompson’s original study of 500 redheads published in 1988, with its statistical analysis performed by Gary Antonacci, found 136 (27.2%) of the redheads were born with Mars within 30 degrees of the Ascendant. This was significantly higher than the control groups, which ranged from 16.9% to 19.4%. The probability of this distribution occurring by chance is less than on in a million (P<0.000001). Mars was also found to have lower frequency (9.8%) within 30 degrees of the Descendant for the redheads compared to the control groups, which ranged from 14.0% to 18.6%. The probability of this result was less than one in 350,000 (P<0.000035). The test data supported both experimental hypotheses regarding higher Mars frequency rising and lower frequency setting.
Although the original experiment included control groups, which would control for astronomical and demographic artifacts, the results were criticized by Geoffrey Dean (Dean, 1988) and challenged by Françoise Gauquelin (Schneider-Gauquelin, 1988). The most critical argument that could potentially explain the findings concerned the nycthemeral “Mars-dawn factor.” This artifact arises from the slight demographic preference for babies to be born near sunrise and the slight astronomical tendency of Mars to be situated near the Sun in the sky. In response to these criticisms, Hill and Thompson submitted the experimental data and the Mars-dawn question for independent testing by Beverly Steffert Ph.D. and Michael O’Neill B.Sc. This testing verified the accuracy of the claimed results and demonstrated that the Mars-dawn factor contributed a Mars variation of less than one percent, which did not explain the experimental findings (Hill; 1988/89, 7).
An additional test performed by O’Neill judged for any unknown demographic or
astronomical bias by using a data shuffling methodology. Data shuffling is typically used in experiments without a control group and its use in this evaluation was somewhat extraordinary. This method used the same distribution as the redheads’ births, with the same year, month, day, place and time of birth, but shuffled
the times of birth. Each redhead kept their real birth date and place, but was
given the birth time of the preceding redhead according to alphabetical order.
Thus, the shuffled test group had exactly the same demographic and astronomical
data as the redheads' group with the real birth times.
O’Neill repeated the counter-test numerous times, each time shuffling farther back in the order. In each test he obtained results similar to the original controls. The Mars distributions for the counter-tests significantly differed from the distribution of Mars for the
real birth times of the redheads. Based on this scrutiny of the data, Steffert and O’Neill found “no case for rejecting the authors’ claim” (Hill; 1996, 32). They upheld the hypotheses but urged replication as a next step by using a new sample of redheads, preferably from another country. They suggested including the charts of 100 redheads collected some 20 years earlier by the late astrologer John Addey. This collection was unknown to Hill and Thompson.
The replications used two new groups of redheads, American and Canadian redheads (N=479) collected by various astrologers, and British redheads (N=473) collected by Steffert and O'Neill. All the evaluations were performed by O’Neill. Two control groups were provided by US researcher Mark Urban-Lurain. One control group substituted Ascendants based on the mean latitude of each group and the other used multiple shuffled birth times.
The replication results for Mars within 30 degrees the Ascendant for the American/Canadian redheads were significant compared to the two controls (P=0.0099 and P=0.048) although these results were noticeably less significant than those obtained from the original experiment. Curiously, the British results for Mars near the Ascendant were evaluated separately in two parts (N=373 new redheads, and N=100 Addey redheads). Though each of the two British result went in the direction of the hypothesis, neither achieved significance. This would seem to suggest that the sample sizes were too small. Smaller sample sizes have the mathematical effect of reducing probability values.
When all the replication groups, American/Canadian and British, were combined (N=952), the results were significant compared to the control groups (P=0.007 and P=0.015), but still less significant than the original experiment. The test for natal Mars within 30 degrees of the Descendant, hypothesized as a negative indicator for redheads, failed to obtain significant results for any of the replication groups, though the results were consistently in the direction of the hypothesis. Again, this result would seem to suggest that the sample sizes were too small.
Based on the directions of the redhead replication data for the Ascendant and Descendant results, Hill argued that further evaluation that takes into account the differences between the Ascendant and Descendant distributions would elevate the probability values in support of the two redhead hypotheses (Hill; 1996, 35). Unfortunately, she did not provide these calculations, which judging from the graphs of data, would seem very likely to lend greater support to her claims.
The Hill/Thompson redheads study is important because it is the first to
discover a significant connection between astrological patterns and a physically
expressed genetic trait. The effect size is small but given the finding of significance, improvements to the experimental methods might be able to isolate key contributing factors.
In regard to the published Hill/Thompson data, one could argue that some of the discrepancy between the original and replication results might have been due to less rigorous sampling methods in the replication.
The initial study included members of redheads clubs, which are more
likely to have emphasized the traits required by the study. The first study also
attempted to rate the tone of the red hair (auburn, strawberry blond,
bright orange, or bright red), which could have produced a scale of red intensity that could be correlated to Mars frequencies. This appears not to have been evaluated, perhaps due to the smaller sample sizes of each tone. Strawberry blonds, who have a less intense tone of red color,
were withheld from the original study but then added into the replication. This might have reduced the significance. From observations in astrological practice outside of the Hill/Thompson study, it is quite possible that dyed red hair could also be associated with natal configurations where Mars is particularly emphasized. This suggests that people with dyed red hair should also be studied as a separate category.
What is especially interesting about the Hill/Thompson finding is that it suggests the distinct possibility of experiments that fully leverage known biochemical and genetic factors. Red hair requires elevated levels of trichosiderin, and this can be quantified from hair samples. The scale of trichosiderin levels could hypothetically be correlated to Mars frequencies. Also, genetic factors for red hair are known and these factors can be explicitly evaluated from DNA samples provided by test participants. In general, DNA indicators appear to have some of the same propensities and complexities commonly attributed to astrological signatures in that the traits they carry are not always expressed in the same way and there can also be numerous related contributing configurations. Consideration of both DNA and astrology together in a cross-disciplinary study of red hair could potentially begin to answer many as-yet unresolved questions in both fields.
Dean, Geoffrey, Letters to the Editor, FAA Journal, October 1988.
Hill, Judith A. and Jacalyn Thompson, The Mars-Redhead Link: A Scientific Test of Astrology, NCGR Journal, Winter 1988-89.
Hill, Judith A., Redheads and Mars: A Scientific Testimony, The Mountain Astrologer, May 1996, pp. 29-40.
Hill, Judith A., Response to Geoffrey Dean’s Critique of “The Mars-Redhead Link,” Above & Below: Journal of Astrological Studies, Winter/Spring 1988/89, pp. 7-9.
O’Neil, Mike, Time-Switching Control Applied to Hill and Thompson's Redhead Data, Correlation, June 1991.
Schneider-Gauquelin, Françoise, Judith Hill and Jacalyn Thompson’s Redhead Research, Astro-Psychological Problems 6.3 (1988) pp. 39-42.
© 2013 Kenneth McRitchie