TL;DR: Some surveys note a correlation between Sun sign and levels of extraversion, but that correlation only shows up in groups of people who already know the traits astrology assigns to their sign. When no one prompts them about astrology, the differences usually disappear. That means there is a real effect, but it is the effect of belief on how people describe themselves, not evidence that celestial bodies shape personality. The works Eysenck co-wrote with Gauquelin, in terms of value, should only be seen as historical documents recording how the debate unfolded, and cannot be treated as evidence confirming validity.
Short answer: the correlation is real, but where does it come from?
In some personality surveys, people born under signs described as "fire" or "air" sometimes rate themselves as more extroverted than people born under other signs. But on closer inspection, that difference appears only among people who already know what astrology says about their sign. When the question gives no hint of astrology, or when respondents are unfamiliar with those descriptions, the difference usually disappears. This is a real correlation in the data, but the cause does not lie in the heavenly bodies.
What is the self-attribution effect?
Self-attribution is the practice of using descriptions already available in a culture to understand and narrate ourselves. People do not invent a personality out of nothing: as we grow up, we hear friends, family, books, newspapers and social media talk about "Aries people", "Cancer people". Those descriptions become part of the vocabulary we use to observe ourselves: "I'm hot-tempered, probably because I'm an Aries". Each time we describe ourselves that way, we both register a trait and reinforce that view of ourselves.
It should be said clearly: this is not a matter of someone deliberately lying on a questionnaire. Respondents answer honestly; they are simply answering from a ready-made frame of reference. This mechanism applies to believers and non-believers in astrology alike, as long as they have heard the popular descriptions at some point.
Why did the 1978 survey attract attention?
In 1978, J. Mayo and colleagues published a survey in The Journal of Social Psychology, reporting a correlation between astrological factors and certain personality traits, including extroversion. The result attracted attention because it ran against the usual expectation of personality researchers: if astrology has no plausible mechanism of action, there should be no significant correlation at all.
That result should be placed in proper proportion: it is a survey, not an experiment controlling for belief. And by this article's convention, a correlation in one sample does not by itself reveal a mechanism. Mayo and colleagues recorded the phenomenon, while the question of "why" remained open.
In the same year, A. G. Smithers and H. J. Cooper published a study in the same journal on personality and birth season, noting a seasonal distribution pattern similar to Mayo's result. These two authors only described the phenomenon; they did not claim it was evidence for astrology.
What did the follow-up checks say?
K. Pawlik and L. Buse re-examined the question in 1979 in Zeitschrift für Sozialpsychologie and found the key point: the correlation appeared only among people who already knew the traits astrology assigns to their sign. The group unfamiliar with those descriptions produced no similar correlation. This is precisely the signature of self-attribution.
Marlis Jackson and Martin S. Fiebert published in 1980 in The Journal of Psychology, finding no correlation between extroversion and Sun sign across 174 students.
Jan J. F. van Rooij published in 1994 in Personality and Individual Differences, showing that the effect appeared only among people who understood astrology, and grew stronger when respondents were hinted that the study related to astrology. The detail "grew stronger when hinted" matters greatly: it shows the difference is not a biological constant, but a context-dependent phenomenon.
In 1988, Jan J. F. van Rooij, Marianne A. Brak and Jacques J. F. Commandeur published a short study in The Journal of Psychology testing the relationship between extroversion and Sun sign, continuing the debate around the results of Mayo, White and Eysenck, and framing it within the self-attribution hypothesis. Only the approach and the context of debate are noted here, without specific figures.
Catherine S. Fichten and Betty Sunerton published in 1983 in The Journal of Psychology, noting that knowing one's sign in advance led readers to find personality descriptions more accurate. This is another piece of the same mechanism: once a frame is in place, people readily find a description "true of me".
| Study | Data scope | Main result |
|---|---|---|
| Mayo and colleagues, 1978 | Survey, The Journal of Social Psychology | Reported a correlation between astrological factors and personality, including extroversion |
| Smithers and Cooper, 1978 | Same journal | Seasonal birth distribution pattern similar to Mayo |
| Pawlik and Buse, 1979 | Re-examination | Correlation only among people who knew the traits assigned to their sign |
| Jackson and Fiebert, 1980 | 174 students | No correlation found between extroversion and Sun sign |
| Fichten and Sunerton, 1983 | Readers of personality descriptions | Knowing one's sign in advance made descriptions seem more accurate |
| van Rooij and colleagues, 1988 | Short study, The Journal of Psychology | Tested the extroversion and Sun sign relationship within the self-attribution frame |
| van Rooij, 1994 | Re-examination | Effect only among people who understood astrology, stronger when hinted |
Why do some surveys still find "fire" signs more extroverted?
Because the survey sample cannot separate the group that "already knows the descriptions of its sign" from the rest. If most participants have read horoscopes in newspapers, or heard friends say "Aries people are hot-tempered", their answers are already tinged with that knowledge. The correlation shows up not because the Sun was shining at the moment of birth, but because culture has assigned the label first, and then people assign that label to themselves.
Some astrological schools hold that these descriptions reflect a real cosmic order, and that people finding them accurate is simply confirmation of that order. The rule of this article is: a system's own texts only demonstrate what that system says, not that what it says is true. The survey results above cannot adjudicate that cosmological question; they only show that most of the observed difference can be explained by psychological mechanisms.
One methodological point deserves note: according to Gerhard Mayer's overview in the Journal of Scientific Exploration, prior astrological knowledge can bias answers, including simple knowledge skimmed from newspapers or packaging. Mayer also reminds us that the Barnum effect, the tendency to find vague descriptions "true of oneself", is one of the well-established psychological mechanisms in this field.
Can Eysenck's writing on astrology be used as evidence?
No. H. J. Eysenck had an article, "Biography in the Service of Science: A Look at Astrology", published in the journal Biography in 1979, during the period of his collaboration with Gauquelin. That article discussed using biographical data to test the link between personality and astrological prediction. But Nhật Nguyệt's record places this article at evidence level D, meaning it has value only as a historical document of the debate, not as evidence for the validity of astrology. In other words, it shows how Eysenck framed the question at the time, but does not prove that the zodiac can measure personality.
How does this apply to "people born in the Year of the Tiger are hot-tempered"?
In folk belief, people born in the Year of the Tiger are described as strong, hot-tempered, fond of leading. The mechanism here is exactly the same as with the zodiac: people hear that description from childhood, and then when observing themselves, moments of anger are remembered more vividly, while calm moments are easily overlooked. Gradually, they narrate themselves using precisely the vocabulary their culture has supplied.
There is no self-attribution research on the zodiac animals or the Five Elements destiny among Vietnamese people. Saying "the mechanism is the same" here is a reasonable hypothesis based on studies of the Western zodiac, not a result measured on Vietnamese people. To know for certain, a survey would be needed that separates the group already familiar with the description of its zodiac animal from the group that has never heard it, and compares against an independent personality scale.
It should also be said clearly: the studies cited above were all conducted abroad, mainly on students and adults in the West, during the 1970s to 1990s. Those results do not automatically carry over to Vietnamese people, and do not automatically carry over to systems such as Tử Vi (Vietnamese astrology) or Bát Tự (BaZi), which use entirely different methods of charting and interpretation from Western astrology.
A real correlation can still have another cause
This is a general lesson, not one specific to astrology. A correlation figure in data only says that two things change together; it does not say which causes which, nor that there is no third thing behind them. Here, the third thing may well be cultural knowledge about the zodiac. When that knowledge is removed from the equation, most of the difference disappears.
The method commonly used in psychology to tell the difference is to separate groups by level of astrological knowledge before asking. If the difference remains only in the knowledgeable group, we are looking at an effect of belief. If the difference remains intact in the group that knows nothing about astrology, then another explanation is needed. So far, published data lean toward the first possibility.
So should we say astrology is "wrong"?
No. The question "can astrology measure personality" is a narrow one, and the answer based on existing data is: there is no evidence that Sun sign predicts personality once the respondent's knowledge is controlled for. But that does not deny that many people find these descriptions useful, that reading about one's sign can be an enjoyable way to reflect, or that a practising astrologer can offer helpful conversations. The self-attribution effect and the Barnum effect are real; they explain why the experience is so persuasive, without needing the hypothesis that heavenly bodies act on personality.
In other words, the distinction worth keeping is this: the subjective experience is real, the mechanism producing that experience is psychological and cultural, and the metaphysical question about the cosmos lies beyond what existing data can answer.
