Astrology testing: birth charts, hidden identities and a misleading correlation
Astrology testing has produced a different kind of record from a horoscope: predictions compared with outcomes under specified conditions. Shawn Carlson's 1985 Nature publication described two double-blind tests of whether natal charts could accurately describe personality. A later experiment by Alyssa Wyman and Stuart Vyse compared astrological descriptions with a conventional personality inventory. A separate statistical investigation examined an apparent connection between zodiac signs and survival in a medical registry. These investigations tested defined claims and methods; none established a reliable planetary influence directing the lives of the people studied.
Carlson's publication identifies its subject as natal astrology, the proposed connection between a person's birth chart and personality traits. The chart represents celestial positions for a particular birth. The Nature record dates the publication to December 5, 1985. Its stated experimental target was accurate personality description, rather than the astronomical calculation of where a planet would appear. That distinction gives astrology testing a measurable task. Celestial coordinates can be calculated accurately while the proposed human characteristics associated with them still fail to match the people concerned.
A contemporary report in the Spring 1986 Skeptical Inquirer preserved Carlson's description of the astrology testing procedure. Astrologers received a natal chart and three anonymous personality profiles, one belonging to the person whose birth data generated the chart. The other profiles came from different people. They had no face-to-face contact with the subjects. The profiles came from a standardized personality inventory known as the CPI. The double-blind arrangement kept both astrologers and testers from knowing the correct matches during the test, reducing opportunities for personal impressions or inadvertent cues to substitute for chart information.
The same contemporary report records the performance benchmark and result. Random selection among three profiles would succeed about one-third of the time. The astrologers expected at least one-half of their selections to succeed. Their actual matching stayed at the chance level, and greater expressed confidence did not bring greater accuracy. The report also records that astrologers contributed suggestions to the design and were familiarized with the personality inventory. Astrology testing here involved a procedure developed with practitioner input, with an expected success rate stated before the outcome was assessed.
Wyman and Vyse's experiment introduced another comparison. In his 2018 retrospective, Vyse described college students choosing between their own computer-generated natal-chart interpretation and one randomly selected from another participant. Chance performance was 50 percent. The students identified their own astrological interpretation 46 percent of the time, a result not significantly different from chance. When offered their own conventional personality profile alongside another person's, they identified it 79 percent of the time. The contrast made the ability to recognize a personal description part of the experiment, instead of assuming every identification task was equally difficult.
Vyse identified the work as the study published with Wyman in the Journal of General Psychology in 2008. Its two-choice design differed from Carlson's three-profile matching by astrologers. The results cannot be combined as though they were the same task with the same participants. They do, however, constrain a shared claim: that a description tied to birth information reliably identifies the individual concerned. Astrology testing in these settings measured correspondence between particular people and particular profiles, using alternatives that were plausible enough to make selection meaningful.
Another approach to astrology testing emerged from an investigation by Alberto Carmona-Bayonas and colleagues. It showed why an apparently positive statistical relationship also needs scrutiny. In the AGAMENON-SEOM registry, zodiac signs initially appeared associated with survival among patients with advanced gastric cancer. Their paper reported a probability value of 0.01 in an initial analysis. Changing the reference sign or regrouping signs produced further apparently significant comparisons. The researchers then modeled seasonality and birth year as potential confounders. After those adjustments, the astrological association no longer remained significant in their final model. Their proposed seasonal pathways were hypotheses about ordinary influences associated with calendar time.
The registry investigation was an observational case study, rather than an experiment assigning people different birth signs. Its authors emphasized that a sign also partitions the calendar and that statistical fit alone does not identify a cause. They noted that some measures of predictive performance could favor a model with a less plausible causal structure. Astrology testing therefore faces more than a contest between a small probability value and a large one. The statistical model must distinguish a purported celestial effect from other variables that travel with a birth date.
Planetary motion itself remains an astronomical phenomenon. NASA's StarChild explanation describes retrograde motion as an apparent reversal against the stars caused by the changing positions of Earth and another planet as both orbit the Sun. Its diagram follows Earth overtaking Mars: the projected direction changes although Mars keeps moving along its orbit. Mercury's apparent reversal likewise does not mean the planet physically reverses its orbit. This geometry supplies no demonstrated link to personal events. Accurate descriptions of the sky and reliable predictions about individual lives require different evidence, and the matching experiments and registry analysis did not supply the latter.
- Carmona-Bayonas et al., Causal considerations can inform the interpretation of surprising associations in medical registries, original research manuscript (2021; journal issue 2022) — pmc.ncbi.nlm.nih.gov
- Shawn Carlson, A double-blind test of astrology, Nature 318, 419–425 (1985), publication record and experimental statement — www.nature.com
- Skeptical Inquirer, Spring 1986, pages 194–196, contemporary report of Carlson’s experiment and his conclusions — skepticalinquirer.org
- Stuart Vyse, Why Are Millennials Turning to Astrology?, 2018: researcher’s description of the Wyman–Vyse experiment — skepticalinquirer.org
- NASA Goddard StarChild, What does it mean for Mercury to be in retrograde?, June 2002, orbital-geometry diagram — starchild.gsfc.nasa.gov