Can Astrology Predict Personality? What the Double-Blind Tests Say

Does your chart really understand you?

Many people read a Tử Vi chart or a birth chart and feel it hits home: the tendency to worry, the stubborn streak, the pattern they keep running into in relationships. That feeling is real, and the psychological question is not whether you feel it fits, but this: if an astrologer is asked to match charts to real people without knowing who is who, do they do better than rolling dice?

That is the question double-blind tests are designed to answer. In this kind of test, the person doing the matching does not know which chart belongs to whom, and the person being read does not know which description is their own. If a chart truly carries information about personality, the correct match rate must rise above chance by a measurable amount. Published tests generally do not show that.

The 1985 Carlson test in Nature: a design approved by astrologers themselves

The best-known study in this line is by Shawn Carlson, published in Nature in 1985 under the title "A Double Blind Test of Astrology". What makes it widely cited is not only the result but how it was prepared: the design was shown to representatives of the astrological community and agreed before the test ran, to avoid later claims that the test had been rigged against astrology.

In the test, astrologers were asked to match charts to participants' California Psychological Inventory (CPI) profiles. The result: the correct match rate was no higher than chance. In other words, without knowing who is who, reading a chart gave no information that distinguished one person from another better than guessing.

An equally important part of Carlson 1985 tested the other side: could each participant recognise their own chart reading when it was mixed in with others? That result also did not exceed chance. This matters because it separates two questions that are often merged: whether the astrologer guesses right, and whether the person recognises themselves.

Do people recognise their own chart?

This question was independently retested more than two decades later by Alyssa Jayne Wyman and Stuart Vyse, published in 2008 in The Journal of General Psychology under the title "Science Versus the Stars". The team used the NEO-FFI (a short form of the Big Five personality model) and computer-generated birth charts. The result split cleanly in two: participants recognised their real NEO-FFI profile above chance, but did not recognise their real astrological reading.

This is a tidy methodological contrast. The same group of people, the same self-recognition task, but when the description was built from psychological test data, people recognised themselves; when it was built from an astrological chart, they did not. If charts carried real personality information, this difference would be hard to explain.

A note on scope: both Carlson 1985 and Wyman and Vyse 2008 were run in the West, with English-speaking participants, Western astrology and Western personality tests. The results do not automatically apply to Vietnamese people or to Eastern numerological systems.

What does the large-scale data say?

Peter Hartmann and colleagues published a large study in 2006 in Personality and Individual Differences, on the relationship between birth date and individual differences in personality and intelligence. The study used two large datasets, one of them with roughly 15,000 people, and checked whether astrological predictions about personality and intelligence by birth date showed up. The result: none of the astrological predictions exceeded chance.

Earlier, Dave Clarke and colleagues published a 1996 study in The Journal of Psychology on the zodiac sign as a determinant of extraversion and emotionality. Notably, this group did not only test the Sun sign; they cast full charts including the Moon and the ascendant, answering in advance the familiar objection that "the Sun sign is not enough, you have to look at the whole chart". The result still showed no relationship as astrology predicts.

This picture is not a few isolated results. Geoffrey Dean and Ivan W. Kelly reviewed astrologer tests from the 1950s onward and found consistently negative results. The methodological point is that such a review does not just gather individual studies; it shows a trend running across decades, research groups and designs.

Why do so many people still find their chart "accurate"?

In folk belief and everyday experience, people often tell each other that a chart "nailed it", "got my personality exactly". This is real for the person experiencing it, but it is not evidence that the chart carries personality information. The feeling of being accurately described can come from several psychological mechanisms, and those mechanisms do not need the chart to be correct in order to operate.

One real effect does arise from belief: for example, believing you are an extravert can lead you to seek out social situations more, and over time behaviour changes and outcomes change with it. This kind of effect is real, but it is not the chart correctly predicting personality; it is belief changing behaviour, and behaviour changing outcomes. Keeping the two apart matters, because merging them leads to wrong conclusions in both directions: either denying a real psychological effect, or attributing to the chart a power it does not have.

On the side of practitioners and researchers of astrology, there is another reading of these tests. Gerhard A. Mayer, in "Astrology and Science: A Precarious Relationship. Part 2", published in 2020 in the Journal of Scientific Exploration, catalogues many methodological problems that could make tests fail. He notes that in matching tests, the astrologer must translate an abstract structure into concrete descriptions of life, and that this translation requires creativity, is easy to get wrong, and is easily distorted by a poor writer. He also cites a matching test in India where the team was asked to distinguish the charts of gifted students from those of students with developmental delay, and according to Mayer the classification result was worse than dice.

This is a critique worth reading, and it belongs to evidence level B: a view sympathetic to anomalous research, raising methodological problems, but not yet producing a new test that replicates a positive result. Mayer also acknowledges that he once took part in a small test where the result matched the participant's chart, but with a sample size of one that counts for nothing.

Where it is still debated: Ertel's reanalysis

Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsThere is a reanalysis of Carlson's data by Suitbert Ertel claiming a small effect that the original design could not separate out. This is a minority view: it has not been independently replicated by another research group, and mainstream researchers generally keep the negative conclusion from the original data. This article mentions it for transparency, not to place it on equal footing with replicated double-blind results.

On the astrological side, there is another explanation for the occasional correct match: that the factor producing the result is not the chart but the astrologer's extrasensory ability. Geoffrey Dean, who once practised astrology and later moved to the sceptical side, wrote of Clark's successful blind tests that they do not prove astrology works, only that the astrologer works. This reading shifts the question to another field, and it too lacks solid evidence.

What this article does not say

This article addresses one narrow question: when astrologers match charts to real people without knowing who is who, do they do better than chance? The answer from the best-controlled tests is no.

That does not say astrology is culturally meaningless, does not say it is useless as a tool for self-reflection, and says nothing about believers or practitioners. Someone sitting down with their own chart and finding more language to think about themselves is entirely possible, and that value is not overturned by these tests. The test answers only the question of the accuracy of the prediction, not the question of the value of the experience.

Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsOn Eastern systems, the evidence gap must be stated plainly: no double-blind test has been run on Vietnamese people, and no equivalent test for Tử Vi Đẩu Số or Bát Tự exists in the literature. This article therefore does not carry Western astrology results over to Eastern Tử Vi. These are two systems with different theoretical frameworks, chart construction and interpretation, and the fact that they have not been tested by the same kind of test is a gap in the literature, not a conclusion. The discussion of the right framing for Eastern systems is set out in the article Tử vi là khoa học hay mê tín: câu hỏi đặt sai chỗ, và câu hỏi nên đặt thay vào đó.