The evidence so far
Has anyone actually tested astrology? Yes. Mostly a different astrology.
A relative has forwarded today's Sun sign column to the family group. Another has replied with a link titled SCIENCE DESTROYS ASTROLOGY. Neither has read past the headline. Both are typing.
"Has astrology been tested?" is usually asked by someone who already has the answer they would like. Half the room is hoping for "yes, and it failed". The other half is hoping for "science can't measure these things". Both halves are about to be mildly inconvenienced.
It has been tested. In the journal Nature, in a classroom, on 173,709 people in China, and on three decades of Swedish marriage records, because Sweden writes everything down. The results are clear. The small print is about what exactly went into the test.
Below is the record: the well-known studies, what each one did, what none of them did, and what Calmistry has promised to do about the gap. We are running a study on this ourselves, so treat us as an interested party and check our working.
The tests, in order of appearance
Six studies carry most of this conversation. Here is what each one did.
- 1949: one sketch for all
A psychologist named Forer gave his students a personality sketch. Each believed it was theirs alone. Everyone had the same text, partly lifted from an astrology book, and they rated it a very good description of themselves.
- 1985: the blind match
In a double-blind test, astrologers tried to match people's birth charts to their personality profiles. They did no better than chance. A 2009 reanalysis argued the data showed marginally better than chance and that the test was too small to be sure. That argument is still open.
- 2006: 15,000 birthdays
Hartmann and colleagues checked over 4,000 men and over 11,000 young adults. Date of birth and Sun sign did not relate to personality or general intelligence.
- 2009: the Pune test
Narlikar and colleagues shuffled the charts of 100 academically bright children and 100 children with intellectual disabilities. Astrologers sorted them at about chance level. The test was small, and many astrologers declined to take part.
- 2020: 173,709 adults
Lu and colleagues found Sun sign predicted none of the five broad personality traits and, in 32,878 people, did not predict job performance. This held for believers and non-believers. People still judged each other by sign in dating and hiring.
- 2020: Swedish marriages
Helgertz and Scott went through Swedish register data from 1968 to 2001. Couples whose Sun signs are said to go well together were no more common among marriages and no less likely to divorce.
What those tests did not test
Read the cards again and look at what was fed into the machine. The three largest studies used the Sun sign: one of twelve labels, fixed by the day and month of birth, the kind printed next to the crossword. That is a fair test of crossword-page astrology, and crossword-page astrology failed it, repeatedly, at enormous scale.
A Vedic chart is a different object. It is worked out from the date, the exact time and the place of birth. It uses a different zodiac, the sign rising on the horizon, twelve houses and 27 lunar mansions, and when people in India say "rashi" they usually mean the Moon's sign, which the big studies never looked at. Finding nothing in Swedish Sun signs tells you about Moon signs roughly what testing a bicycle tells you about a ferry. The ferry may also sink. Nobody has put it in the water.
The 1985 and Pune tests did use full charts. They asked whether astrologers, reading by judgment, could pick the right person, and both were small. Neither asked the plainer question: across thousands of people, do specific chart features line up with something measured?
So the scoreboard has three lines. Sun signs against personality: tested hard, nothing found. Astrologers matching charts to people: tested a few times, about chance, one result disputed. Vedic chart features against measured stress patterns in a large sample: no result in either direction, because nobody appears to have run one at scale. An untested claim has not passed anything. It is simply still waiting in the corridor.
The Barnum effect, and our use of it
The Barnum effect is what Forer's classroom showed. A statement that fits nearly everyone feels as if it was written for you. "You can be sociable, but you also need time alone." "You are harder on yourself than people realise." His students were not foolish. The sentences were true. They were just true of everybody on the bus.
Now the awkward part. Many reads on this site carry a list headed something like "if it's at play, you'd probably nod at a few of these". Those lists are built to be nodded at. Most people with the problem would tick most of the lines, whatever their chart contains. That is the Barnum effect, used on purpose, by us. A tick tells you the sentence is relatable. It tells you nothing about your chart.
The same caution applies to the quiz. Being told you came out closest to bamboo is pleasant, and pleasant is exactly the condition under which people stop checking.
The fair-test promise
Since we know the lists flatter, the study is designed so that flattery cannot count as a result. Three commitments do the work.
- We wrote our expectations down first. Before looking at any data, we recorded which chart features should go with which of the thirteen stress properties. Nobody gets to discover afterwards that the result was what we meant all along.
- Real charts against shuffled ones. Each pattern is checked twice: once with every person's own chart, and once with the charts dealt out at random to the wrong people. If a stranger's chart "fits" you as well as your own, the fit was Barnum.
- We publish whichever way it turns out. Something, nothing, or something embarrassing for us.
Who is invited
Everyone, openly. The study needs people who expect it to work, people who expect it to fail, and people who only came for the quiz. A test that only sceptics sit is as lopsided as one that only believers sit.
Until it reports, the accurate summary is short and satisfies nobody. Magazine Sun signs have been tested thoroughly and showed no link to personality. Full Vedic charts have barely been tested at all. Anyone telling you more than that, in either direction, has got ahead of the data.
Quick answers
Has astrology been scientifically tested?
Yes. Large studies in 2006 and 2020 found Sun signs did not predict personality, intelligence, job performance or divorce. A 1985 double-blind test found astrologers matched charts to personality profiles no better than chance, though a later reanalysis disputes how firm that result is.
What is the Barnum effect?
It is the tendency to rate a general description that fits almost anyone as a very good description of yourself. Forer showed it in 1949 by giving every student the same personality sketch.
Has Vedic astrology been tested?
Barely. A small 2009 test in Pune found astrologers sorted children's charts at about chance level. We know of no large test of Vedic chart features such as Moon sign or nakshatra against measured personality or stress. Calmistry's study is an attempt at one.