more details - Carlson 1985

Purpose

To test the “fundamental thesis of natal astrology” i.e. that “the positions of the planets …. at the moment of birth can be used to determine the subject’s general personality traits and tendencies in temperament and behaviour, and to indicate the major issues which the subject is likely to encounter”  with the co-operation of scientists, statisticians and astrologers.

Methods

Scientific and statistical advice was provided by a professor of Physics working at the University of California.

The National Council for Geocosmic Research (NCGR) provided a list of 90 astrologers who had some background in psychology and were familiar with psychological tests. 

The subjects were recruited by means of advertisements in the San Francisco Bay area newspapers, classroom announcements and postings on the Berkeley campus.  Approx. 70% were college students; they all had to be at least 17 years old.  They were  asked about their belief in astrology and whether they already had had their natal chart drawn up. Subjects who already had had their charts drawn up and those who were strongly against astrology were eliminated in the attempt to avoid bias. 

The subjects were subdivided into two groups, the test subjects and control subjects.  The test subjects were matched to control subjects, who were to belong to the same sun sign (many people know the main features of their sun sign and might use these to identify the correct interpretation) but had also to be at least 3 years older or younger than the test subject, so that their natal chart would be significantly different. 

Birth information:  all the subjects filled in a form providing birthday, time and location of birth. Birth time had to be accurate to within 15 minutes and had to be supported by documentation (birth certificate, hospital record, county record, a “baby book” provided that the birth time had been entered when the child was born).

Natal charts were calculated by NCGR on the basis of the provided birth data using the Digicomp DR 70 Astrological Computer and checked by manual calculation.  The natal charts were coded and then passed on to the astrologers who wrote an astrological interpretation.  The interpretation had to comply with a standard format so that it would be as uniform as possible and would not provide any clues that could help subjects to identify the right interpretation.

Participating subjects were asked to reply to the 480 true-false questions of the California Personality Inventory (CPI), which is a standard personality test that ranks subjects in terms of 18 personality attributes. The test was chosen over other psychological tests by the astrologers, who considered it to be the most suitable for the investigation.  The tests were graded by volunteer undergraduate students who were not connected to the study in any other way; scores were then plotted on a graph that illustrates the personality profile.

Sample size:  the originally planned sample size was 256 – 128 assigned to the test group and 128 to the control group.  Unfortunately, many subjects dropped out.  Only 177 subjects completed test no 1A:  83 test subjects and 94 controls.  Even fewer subjects completed test 1B:  116, 56 test subjects and 50 controls.  Only 28 astrologers were available for test 2. 

Blinding:  An assistant of the Professor of Physics assigned a code number to each subject and was responsible for data files and giving out anonymous study material bearing only the code number.  CPI profiles were relabeled with another code calculated according to a pre-set formula, so that subjects and astrologers could not match them to natal charts or the astrological interpretations. 

Test No 1A  - Volunteers selecting their own astrological interpretation

Each test subject was given three anonymized interpretations, one of which was his/her own and attempted to select the right one.  They were allowed to rank the three interpretations i.e. to have a first and second choice,  and were also asked to assess how well the interpretation fit them on a scale from 1-10 (10 being 100% fit).

The control group was given the same task.  Each control subject received the same three anonymized interpretations as the matching test subject.

According to the laws of chance, each test subject had a 33% probability of selecting the right interpretation.  It was agreed that if the volunteers selected the right chart at least 50% of the time, the data would be considered to support the validity of astrology.

Responses were received from 83 test subjects and 94 controls.  First choice selections were correct in 33.7% and 44.7% of cases, respectively.  Adding the second choice, the correct response rates increased to 73.5% and 80.8%.  The conclusion was that the performance of the test subjects was not better than chance.

Test no 1B - Volunteers selecting their CPI profile

Each test subject was given three CPI profiles together with general explanatory material on the test.  One of the CPI profiles was his/her own and the subject was to attempt to select the right one.  The same three CPI profiles were given to the matching subjects in the control group.

Responses were received from 56 test subjects and 50 controls.  First choice selections were correct in 44.6% and 42% of cases, respectively.  Adding the second choice, the correct response rates increased up to 73.2% and 68.0%, respectively.  The conclusion was that the performance of the test subjects was not better than chance.

Test No 2 – astrologers selecting the corresponding CPI profile

Each participating astrologer was given a pre-agreed number of natal charts (usually 4) together with 3 CPI profiles related to each natal chart.  No information on the subject was provided, not even gender.  The astrologers attempted to select the right CPI profile out of the three provided for each natal chart.  In order to save the astrologers work, they were allowed to make the CPI matchings to the natal charts they had already interpreted.  They were allowed to rank the three CPI profiles i.e. to have a first and second choice and were also asked to rate how closely the CPI profile fit the natal chart.  

According to the laws of chance, each astrologer had a 33% probability of selecting the right CPI profile.  It was agreed that if the astrologers selected the right chart at least 50% of the time, the data would be considered to support the validity of astrology.

The first choice was correct in only 34.4% of cases.  When the second choice was taken into consideration the correct answers were provided only in 39.6% of cases – less than the probability according to a random selection (66%).   Results did not improve when the rating of the level of fit was taken into consideration. Therefore, also the performance of the astrologers was not better than chance.

Consequently, the working hypothesis that the fundamental thesis of natal astrology is true was rejected as untrue.

COMMENTS

Do the results of Test no 1 prove that astrology is FALSE?

NO, they don’t

The researchers themselves came to the conclusion that the test does not prove anything.  The fact that the subjects were unable not only to identify their astrological profile, but also their psychological profile, suggests that people have poor insight into themselves and cannot an recognize accurate description of their personality, or at the very least, that they have serious difficult in matching the CPI profile to their knowledge of themselves. 

In other words, this type of test is not suitable for the assessment of the validity of astrology

Do the results of Test no 2 prove that astrology is FALSE?

The results of Test no 2 are not in favor of astrology, but …

This study was performed 35 years ago.  Today I do not think that any astrologer would accept to participate in an investigation where birth times are provided with a broad approximation that could have a major impact on results (15 minutes).  In this study the birth time was based on a birth certificate, a hospital record or even a “baby book”.  Unless the reporter was interested in astrology and knew how important the accuracy of a birth time is up to approx. one (yes, one!) minute, the information on such documents is inaccurate. A technique called rectification has been developed by astrologers, whereby they correct the birth time based on important life events, such as marriage etc.  The problem is that the technique is tedious and takes a long time, often the best part of a day, so funding should be available for the correction of each birth time by a separate astrologer and the study would become expensive.

The CPI profile may have been an inappropriate choice.  The factors that prevented the subjects from identifying their own profile may have also prevented the astrologers from selecting the right profile (e.g. difficulty in interpreting the data provided).  Moreover, the CPI profile does not include key data that might have enabled the astrologers to achieve correct matches, such as important life events (marriage, etc) that astrologers routinely predict during consultations. 

According to me, the results are therefore to be considered inconclusive. 

Additional criticisms has been made by the Astrologer Geoffrey Cornelius in his Book "The Moment of Astrology".  He points out that the test was carried out in artificial conditions that do not reproduce astrological practice and, more importantly, that the assessment did not regard astrology per se, but rather the capacity of astrologers to interpret natal horoscopes.  Was the sample included in the study representative of all formal astrological interpretations of natal horoscopes by practitioners recognised as competent by NCGR?  There is no answer, since the issue was not even taken into consideration