Carlson's experiment
The Shawn Carlson's double-blind chart matching tests, in which 28 astrologers agreed to match over 100 natal charts to psychological profiles generated by the California Psychological Inventory (CPI) test, is one of the most renowned tests of astrology, and was published in a highly prestigious journal, Nature. Double blinding helps to practically eliminate all bias from a study, including from participants as well as the person performing the study. The experimental protocol used in Carlson's study was agreed to by a group of physicists and astrologers prior to the experiment. Astrologers, nominated by the National Council for Geocosmic Research, acted as the astrological advisors, and helped to ensure, and agreed, that the test was fair. They also chose 26 of the 28 astrologers for the tests, the other 2 being interested astrologers who volunteered afterwards. The astrologers came from Europe and the United States. The astrologers helped to draw up the central proposition of natal astrology to be tested. Published in
Nature in 1985, the study found that predictions based on natal astrology were no better than chance, and that the testing "clearly refutes the astrological hypothesis".
Dean and Kelly
The scientist and former astrologer, Geoffrey Dean and psychologist Ivan Kelly, conducted a large scale scientific test, involving more than one hundred cognitive, behavioural, physical and other variables, but found no support for astrology. A further test involved 45 confident astrologers, with an average of 10 years experience and 160 test subjects (out of an original sample size of 1198 test subjects) who strongly favoured certain characteristics in the Eysenck Personality Questionnaire to extremes. The astrologers performed much worse than merely basing decisions off the individuals age, and much worse than 45 control subjects who did not use birth charts at all.
Other tests
A meta-analysis was conducted, pooling 40 studies consisting of 700 astrologers and over 1,000 birth charts. Ten of the tests, which had a total of 300 participating, involved the astrologers picking the correct chart interpretation out of a number of others that were not the astrologically correct chart interpretation (usually three to five others). When the date and other obvious clues were removed, no significant results were found to suggest there was any preferred chart.
In 10 studies, participants picked horoscopes that they felt were accurate descriptions, with one being the "correct" answer. Again the results were no better than chance.
In a study of 2011 sets of people born within 5 minutes of each other ("time twins") to see if there was any discernible effect, no effect was seen.
Quantitative sociologist David Voas examined the census data for more than 20 million individuals in England and Wales to see if star signs corresponded to marriage arrangements. No effect was seen.