Zodiac Compatibility: What the Research Actually Shows
No — and not for lack of anyone looking hard enough to find it. Astrology's compatibility claim says a birth chart can sort two people before they've spoken; demographer David Voas tested that claim against the birthdays of every married couple in England and Wales — more than ten million marriages — and the sign pairings landed exactly where random chance puts them. That matters most to anyone who has ever ruled someone out because a horoscope called the pairing wrong, or leaned toward someone because it called the pairing right.
Two people can share a sign and never meet. Two people can share nothing on a chart and still be married forty years later. The census doesn't know which one you were rooting for, and it turns out neither does the sky.
Write down what actually held two people together instead of what a chart said should have.
Start writing what actually explains youVoas didn't set out to debunk anything. He noticed that someone else had already commissioned a table of husbands' and wives' star signs from the 2001 census, and planned to run one statistical test, announce the result and provide fifteen minutes of entertainment. The table showed something odd instead: about 22,100 more same-sign marriages than chance would produce, with a spillover toward neighboring signs. That's the number that would have made headlines if he had stopped there.
He didn't stop there. Chasing the pattern down to exact day and month of birth, Voas found the excess was mostly couples who share an actual birthday — most likely a census-form error rather than a romance, since one member of the household usually fills in the form and can write their own birthday in for the spouse. Strip out shared birthdays, shared months, and the census system's habit of entering an illegible or missing birth date as "the first of the month," and what's left is flat: no more same-sign couples than a random shuffle produces.
"In fact, the numbers are just what we'd predict based on chance." — David Voas, demographer, Skeptical Inquirer (2008)
The person who had first requested the data described himself as "a philosopher with a degree in statistics" who had found "subtler, scientifically reliable patterns" in the same numbers (Voas, 2008) — then refused to share them unless Voas agreed in writing not to publish first. Writing down what actually happened between two people, rather than filing it under a category, turns out to be the harder and more honest version of the same project.
Sun signs alone are a blunt target, astrologers reasonably object — a full birth chart, read by a trained professional, is the real test. Physicist Shawn Carlson built exactly that test in 1985, with a protocol that astrologers agreed to before it began. 28 astrologers, working from complete natal charts, tried to match each of more than 100 real subjects to the correct one of three personality profiles generated by a standard psychological inventory, under conditions where neither side knew who was paired with whom.
The astrologers picked correctly at almost exactly the rate you'd expect from guessing. Not a little better, not a little worse — chance. Because astrologers had signed off on the protocol in advance, the result was hard to blame on the wrong test.
Relationship scientists asked a more useful question than which sign goes with which: of the hundreds of variables proposed over a century of research, which ones actually predict how happy two specific people are together? A 2020 collaboration across 29 labs ran machine learning over 43 longitudinal studies of more than 11,000 couples to find out.
Traits didn't win. Life satisfaction, attachment style, depression — the closest scientific cousin to "personality is destiny" — explained a real but modest 21% of relationship quality at the outset. What mattered more, by a wide margin, was how each person felt about the specific relationship they were already in:
- How committed they believed their partner was.
- How appreciated they felt.
- How sexually satisfied they were with their partner.
- How satisfied they believed their partner was.
- How much conflict the relationship carried.
Relationship-specific measures like these five, self-reported and tied to one relationship rather than one personality, predicted up to 45% of the variance in relationship quality at the start of the studies. What no combination of answers could do well was predict who would grow happier or unhappier over time. Knowing the partner's own answers to the same questions added nothing once you already had one person's side. A sign can't move if you switch partners. These numbers can — because they were never about who either of you is in general, which is the part synastry has always tried to skip.
| Zodiac compatibility claim | What the evidence actually shows |
|---|---|
| Certain sign pairings show up more often among married couples | No excess pairings across ten million England and Wales marriages once response error is removed |
| A trained astrologer reading a full chart is the real test | 28 astrologers matched 100+ charts to personality profiles at chance level, double-blind |
| A description feels accurate because it was written for you | An identical sketch from an astrology book was rated 4.3 out of 5 for accuracy, on average |
| Compatibility can be read off two people before they've spoken | How each person feels about their own relationship, not their traits, predicted up to 45% of its quality |
None of that data came from a chart. It came from what people actually said about one specific relationship.
Write the specific thing, not the category
Because the sketch is built to feel specific while staying true of almost everyone. Psychologist Bertram Forer gave a personality test to 39 students in 1949, then handed each of them what he said was their individual result — every student got the same page, lifted wholesale from a newsstand astrology book, and rated its accuracy from 0 to 5. The average came back at 4.3, from a room full of people reading an identical page.
That's not gullibility so much as a description doing its job. "Guarded around new people but fiercely loyal once you commit" reads true for almost anyone who has ever been in a relationship, sign notwithstanding — and that's precisely the appeal a horoscope column or an astrology app understands better than most dating advice manages to. It costs nothing, answers instantly, and never once asks you to say anything real about yourself first.
Anketta doesn't ask what your sign is, doesn't run a quiz to guess how well you'd get along with someone, and has no field for a chart. Beyond hard filters like age and city, a preference model decides which manuscripts reach each other, and it learns from which lines a reader actually highlights — the sentence about wanting kids by thirty, not a moon sign.
That's closer to what the 43-study review found predicts anything at all: a specific fact about this person and this relationship, not a category assigned before either of you learned to talk. Writing the specific thing instead of the category is the harder version of compatibility, and the only one with data behind it. Two people who've actually named what they individually need, rather than compared charts, are working from something a horoscope was never built to measure — and reading someone's actual sentences instead of their sign is where that starts.
Do compatible signs actually get along better in real relationships?
No large study has found this. The biggest test — more than ten million marriages in England and Wales — found sign pairings distributed exactly as random chance predicts, once census recording errors like shared or misremembered birthdays are corrected for.
Does reading a full birth chart work any better than a quick horoscope?
Not in the one rigorous test of the stronger claim. 28 professional astrologers working from complete natal charts matched over 100 subjects to their correct personality profiles at almost exactly the rate guessing would produce, under double-blind conditions designed to rule out other explanations.
Why do so many people say astrology was "eerily accurate" about their relationship?
Because vague, broadly true descriptions feel personally written. In a classic 1949 demonstration, 39 students all received the identical personality sketch, lifted from an astrology book, and rated it 4.3 out of 5 for accuracy on average — a sign the description fit almost anyone, not that it fit them specifically.
Is there anything genuinely useful about astrology in a relationship?
As a shared way to talk about temperament, sure — plenty of couples use it the way others use a favorite quiz, as a starting point for a real conversation. The evidence doesn't support it as a predictor; it hasn't been shown to forecast outcomes the way a couple's own reported feelings do.
What actually forecasts whether a relationship works, if not compatibility on paper?
Across 43 studies of more than 11,000 couples, the strongest predictors of relationship quality were how committed each person believed their partner was, how appreciated they felt, sexual satisfaction, and how much conflict existed — all specific to that one relationship, none of them a trait either partner walked in already carrying. Changes in happiness over time proved largely unpredictable from any of them.
Should two people just avoid mentioning their signs at all?
There's no need to. The mistake isn't mentioning a sign — it's treating it as information the other person hasn't given you yet. Two Geminis who've actually told each other what a hard week costs them have more data than any chart offers, and it's the only version of the data that ever predicted anything.
The chart was never wrong about you, exactly — it just never had anything to be right about. Write the Tuesday down instead.
Unsure about writing? Try reading first.