← Blog
RelationshipsSeptember 20, 2026By Johnson

Where Personality Type Compatibility Charts Actually Come From

Look up any two type nicknames and two sites will give opposite verdicts with equal confidence. The rankings descend from a system Myers and Briggs never wrote, and the largest samples put pairing last.

Somebody looks up Campaigner and Adventurer. The first page says the pairing has natural chemistry and shares a love of spontaneity. The second says both partners avoid structure and the relationship tends to drift. Neither page cites anything.

The pairing is ENFP and ISFP. Both descriptions were written from the same four letters, and the disagreement is not a disagreement about evidence, because there is no evidence in the room.

256

ordered pairings a complete chart has to fill

6%

of the variance in relationship satisfaction explained by a person's own personality

3%

at most explained by their partner's

0.5%

explained by how similar the two of them are

Myers and Briggs never published a compatibility chart

The instrument that produced the four letters reports preferences for one person. Its documentation contains type descriptions and development material, and no ranking of pairings, no list of ideal matches, and no warning about difficult combinations.

The current publisher is blunter about this than most of its critics expect. Its own guidance on relationships says that people are more than four preferences, that personality is extremely complex, and that what contributes to a lasting relationship is a common language for understanding differences rather than a favourable combination of letters.

So the charts came from somewhere else.

The rankings came from a separate system built in Lithuania

Aushra Augustinavichute began work in Vilnius in 1968 on what became Socionics, a typology that starts from the same Jungian material as the four-letter tradition and then adds something the four-letter tradition does not have: a model of how information moves between two people, and a complete set of names for the sixteen ways a pair can relate.

Duality is the most complementary of those relations. Mirror, activity, conflict and twelve others fill out the grid. The point of Socionics was always the grid, which is why it produced pairing verdicts while the Myers tradition produced individual profiles.

The two systems use overlapping notation, so when Socionics material was translated into English and the sociotypes were written as four letters, the rankings traveled with the letters and arrived in a tradition that had never claimed them. Golden pair is a popular rendering of duality. So is the idea that certain pairs are natural opposites who complete each other.

Where a ranking comes fromWho built itWhat it was built on
Duality and the sixteen intertype relationsAushra Augustinavichute from 1968A model of information processing layered onto Jung
The four temperaments and their pairingsDavid KeirseyObservable behaviour grouped into four temperaments
Relationship articles under the modern nicknamesA commercial profile siteIts own five-aspect model built on trait research
The original four-letter instrumentMyers and BriggsNothing, since it published no pairing guidance at all

Read the last row against the first three. Every confident verdict on the internet about which types belong together comes from one of the upper rows, and almost none of them say so.

What the large samples say about pairing

There is a clean test of the underlying idea, and it has been run on samples far larger than anything in the type literature.

Dyrenforth, Kashy, Donnellan and Lucas used three nationally representative samples of married couples in 2010, from Australia, the United Kingdom and Germany, totalling more than 23,000 people. They separated three effects. How satisfied a person is given their own personality, how satisfied they are given their partner's, and how satisfied they are given how similar the two personalities are.

Variance in relationship satisfaction explained across three national samples of married couples
Own personality6%
Partner personality3%
Personality similarity0.5%
  • Own personality: 6%
  • Partner personality: 3%
  • Personality similarity: 0.5%

Similarity came last by an order of magnitude, and it stayed last in all three countries. A compatibility chart is a device for converting the smallest of those three columns into a verdict.

Montoya, Horton and Kirchner found something that explains why the charts still feel right. Pooling 313 investigations in 2008, they reported that measured similarity predicts attraction strongly before two people have met and not at all inside existing relationships, while perceived similarity predicts attraction at every stage. What matters is whether two people feel alike. Whether they are alike, by any instrument, stops mattering once they know each other.

Why any pairing can be narrated both ways

Take any two codes and the same difference supports opposite stories. One partner plans and the other improvises, which is either balance or a standing argument about the weekend. Both partners feel deeply, which is either intimacy or two people with no thermostat in the room.

That flexibility is the reason compatibility content is so easy to produce and so hard to falsify. A verdict that can be written in either direction from the same facts is not making a claim about the world, and the reader supplies the evidence from their own relationship while reading.

What a pairing chart offers
What the question underneath it needs
A verdict about two categories
A description of one specific recurring fight
An explanation assembled after the fact
Something that would have predicted it in advance
A reason the friction is structural and permanent
Whether the friction survives a changed arrangement
Language for a difference
Agreement about what to do with the difference
A ranking out of sixteen
A comparison of two people on the same dimensions

What to do with the question instead

The search that brings people here is almost never idle. It follows a specific incident, and the hope is that a chart will say whether the incident is fixable.

The comparison worth making is dimensional rather than categorical. Two people who both score near the middle of an introversion scale will negotiate a weekend very differently from two people at the far ends, and the four letters hide exactly that difference. The couple compatibility test here scores both partners on the same dimensions and reports where the gaps are largest rather than issuing a verdict on the pairing, which is the only version of this exercise the evidence supports. The comparison between the type frameworks covers which of them are measuring and which are naming, and the ENFP profile keeps the compatibility list on the page while saying plainly what it does not know.

Nobody looks up two type names to learn about typology. They look it up to find out whether the thing that happened last week is a sign. A chart will always answer that question, and the confidence of the answer is the strongest evidence that it was not measured.

Sources

Frequently asked questions

Which type pairings actually divorce less often?

No dataset large enough to answer that question has been published, and the claims circulating about it trace back to small convenience samples rather than to registry data. Divorce research consistently finds its strongest predictors elsewhere, in age at marriage, education, income stability and conflict behaviour. Any page assigning a divorce risk to a four-letter pairing is reporting a theory rather than a measurement.

Is it a bad sign if my partner and I share a type?

It is not a sign of anything on its own, and the popular claim that identical types stagnate has no measurement behind it. Two people with the same four letters can differ sharply on traits that four letters do not capture, particularly conscientiousness and emotional stability, which is why identical codes often feel less alike in practice than the label suggests. The more useful comparison is dimensional, since two people who both came out introverted at 51 percent and 88 percent are not living the same life.

Should a dating app let people filter by type?

Filtering removes people on the basis of a variable with a weak relationship to the outcome, which costs more than it saves. The published work on partner selection finds that the traits people say they want in advance predict their actual attraction to a specific person poorly, so a filter set before meeting anyone is optimising against a preference that will not survive contact. A type filter is best understood as a conversation topic that has been given administrative power.

Why do the golden pair descriptions feel so accurate?

Because they are written to be recognised rather than to be tested, in the same way horoscopes are. A description of a pairing that mentions one partner grounding the other and the other opening the first one up will fit most couples who have been together for a while, since both of those things happen in most relationships. The recognition is evidence about how the text was written and not evidence about the pairing.

Is there any pairing research worth reading?

The literature on similarity is the closest thing to it, and the useful finding is that perceived similarity behaves differently from measured similarity. Couples who believe they are alike report more satisfaction whether or not the questionnaires agree that they are, which suggests the feeling of being understood is doing the work rather than the overlap itself. That points at conversation rather than at matching.

See this pattern in your own numbers

✦

This article, about your situation

Luna has read the same research β€” and, if you let her, your results.

✦ Talk it through

More notes on people