Two INFJs Can Be Nothing Alike. Here's Why
By PsyCove Editorial · 10 min read · Updated 2026-10-27 · Checked against the sources listed below
- A type summarizes four positions, and the positions carry the information. Two people with the same letters can sit near opposite ends of the same dimension.
- The dimensions are continuous, and most people are near the middle. Rescored with a finer method, the two-humped distributions flattened into one center-weighted hump (Bess & Harvey, 2002).
- People move inside their own range. Repeated state reports show a broad band of behavior per person, not a fixed point (Fleeson, 2001).
- Reliability is not the weak link. The indicator's scales show strong internal consistency and test–retest estimates; the problem is what a reliable score can say about one person at one moment (Capraro & Capraro, 2002).
What Do Two People With the Same Type Actually Share?
Two people with the same type share a label that compresses four scores into one word, and what they have in common is exactly the part the label kept.
The type is produced by sorting each of four continuous dimensions into a high half and a low half. Sixteen cells come out. Everyone inside a cell gets the same four letters, and the letters record the side of the midpoint, not the distance from it. The argument would be wrong if people inside one of the sixteen cells resembled each other more than they resembled neighbors; the spread within a cell is too wide for that.
That is the whole mechanism, and it explains the observation without any appeal to maturity, upbringing or life experience. Two people in the same cell can be barely past the line on all four or far beyond it on all four, and the label will look identical in both cases.
Why Does a Shared Label Feel Like Shared Substance?
Because labels are cheap to compare and scores are not, and because the label is the part people repeat.
When someone says they are an INFJ, they are not handing over four numbers. They are handing over a name that four very different people might share. The name survives the conversation because it is short. The numbers behind it never come up.
The compression is worth putting in arithmetic. Four dimensions scored from 0 to 100 allow 100 million distinct profiles. The type system sorts those into 16 boxes, which means each label stands in for roughly six million different profiles. Two people drawn from the same box are, on average, no more similar than two strangers who happen to share a birthday.
How Much Does One Person Move Inside Their Own Type?
One person moves inside their own type more than the language of type suggests. A person is not a point on a dimension; they are a distribution across it.
The clearest evidence comes from studies that ask people to report their current state repeatedly rather than their general self once. Across weeks of repeated reports, each person occupied a wide band on every dimension, and the band was wide enough that the same person could look like two different people on two different days (Fleeson, 2001).
Situations do predictable work on that band. The cognitive-affective system account holds that a person's behavior varies by situation in a stable, if-then pattern: if the setting is one kind, she does this; if it is another, she does that. The variability is part of the personality, not noise around it (Mischel & Shoda, 1995).
Observers notice this too. When strangers rated people from short videotaped episodes, the consistency of the impression across fifteen different situations sat around .43 — substantial, and nowhere near a fixed signature (Borkenau et al., 2004).
Where Does the Four-Letter Score Lose the Detail?
Three places, and none of them is a mistake in the scoring. They are what sorting does.
- The midpoint. Anyone within a few points of the line can be sorted either way by the smallest change in mood or wording (Bess & Harvey, 2002).
- The distance. A score of 55 and a score of 95 produce the same letter. The letter cannot tell you which one you are holding.
- The shape. Four separate dimensions are reported as a single word, so a person who is extreme on one dimension and neutral on three looks the same as someone who is mildly above the line on all four.
Reliability is not where this breaks. A meta-analytic review of the indicator's score reliability found strong internal consistency and test–retest estimates across studies, with variation between them (Capraro & Capraro, 2002). A reliable score can still be an imprecise description of a person, because the reliability is about the instrument repeating itself, not about the person being simple.
What If Two People Differ on Three of the Four Dimensions?
Then they share one letter and almost nothing else, and the conversation about "us INFJs" is doing something other than describing them.
| Dimension | Person A | Person B | Letter shared? |
|---|---|---|---|
| E / I | 62, just over the line | 94, far into it | Yes — I |
| S / N | 58, just over the line | 88, far into it | Yes — N |
| T / F | 54, just over the line | 79, far into it | Yes — F |
| J / P | 52, near the middle | 51, near the middle | Yes — J |
Both people are INFJs. One is mildly introverted, mildly intuitive and mildly feeling. The other is strongly all three. If you met them separately you would probably not guess they shared a category, and you would be right about the impression even though you would be wrong about the letters.
The same logic works across types, which is the more uncomfortable version. Someone at 49 on a dimension and someone at 51 sit in different boxes and are, for practical purposes, the same. The system has a sharp line where the underlying thing has none.
Why Do the Same Descriptions Still Fit Both?
Because type descriptions are written for the box, not for the position, and language written for a box is language most people in it can endorse.
This is the oldest finding in personality assessment. Forer handed students identical feedback and told each of them it was theirs; they rated its accuracy highly. The text was assembled from horoscopes and generalities (Forer, 1949). A description that fits everyone feels like a description that fits you.
Type writing inherits the problem in a specific form. Descriptions are usually written to be recognizable rather than falsifiable, so they lean on statements that most readers can find some evidence for. And the more evaluative a trait is, the more the description drifts toward flattery, which is exactly where observers disagree most (John & Robins, 1993).
“A type description that nobody in the type could disagree with is not a description. It is a compliment with a code name.”
What About the Function Order They Are Supposed to Share?
The stronger version of the claim is not that two INFJs share four letters. It is that they share an order of mental functions, and nothing in a four-dimension score tests that.
A test that sorts four dimensions produces four positions. It does not measure introverted intuition or extraverted feeling, and no arithmetic performed on four letters reconstructs a stack. The argument for keeping those two claims apart is set out in the review of the eight-function model.
- “We are both INFJs” can mean we fall on the same side of four midpoints. True when the letters match, and silent about distance.
- It can mean we process the world the same way. That is a claim about functions, and the score does not address it.
- It can mean we would react alike to the same situation. That is a claim about if-then patterns, which vary inside one person from week to week (Mischel & Shoda, 1995).
The psychometric criticisms land in the same place. A forced choice throws away the information at the cut point, and the resulting categories behave like divisions of a continuous scale rather than natural kinds (Boyle, 1995; Pittenger, 2005). That is a claim about what the instrument can support. Our result page prints a reference type and four positions, and stops there for that reason.
So when two INFJs turn out to be nothing alike, the model has not been refuted. The extra claim was never on the table.
Why Do Type Communities Resist This?
The argument above is unwelcome in type communities for a reason that has little to do with evidence. A shared label does social work: it hands strangers a topic, a vocabulary for their differences, and a place to sit.
Naming that plainly is fairer than treating the resistance as stupidity, because the resistance has three real sources.
- The label is an entrance, not a claim. "We are both INFJs" is a greeting before it is a proposition. Asking someone to drop it asks them to give up the door.
- The description arrived with the community. Most people met their type through a forum, a friend or a post rather than a manual. The fit they felt was produced socially as much as psychometrically.
- Precision sounds like exclusion. "We are on the same side of four cut points" is accurate and cold. Communities are built to be warm, and warmth is not a measurement error.
You can keep the community and drop the metaphysics. The people are real, the vocabulary is useful, and the four letters are a coarse index of where you landed. None of that requires the letters to name a kind.
How Should You Hold Your Own Type, Then?
Use it as the shortest true sentence about you, and keep the rest of the paragraph in view.
- Keep your positions. If you have your four scores, keep them next to the letters. The letters are the index; the numbers are the content.
- Expect to move. Your result describes a range you occupy, not a setting you were assigned. A week that contradicts your type is data, not an error.
- Notice when the description flatters. If every line in your profile reads as a strength, the profile was written to be liked. Half of a fair description should be inconvenient.
- Compare positions, not letters. Two INFJs who are 55 and 95 on the same dimension have more to say to each other about that dimension than about their shared label.
- Refuse the membership test. “You are not really an INFJ then” is a move that protects the category, not a claim about you.
We write the result descriptions in the second person and keep them short, for the reason above. A long description gives a reader more chances to find one line that fits, and fewer chances to notice the lines that do not. Our test of a sentence is whether someone could disagree with it.
Frequently asked questions
Can two people with the same MBTI type be very different?
Yes, and the system predicts it. The type records which side of four midpoints a person fell on, not how far from the midpoint. Two people can share all four letters while sitting at nearly opposite ends of every dimension (Bess & Harvey, 2002).
Does this mean the MBTI is unreliable?
No. Reliability studies of the indicator's scales report strong internal consistency and test–retest estimates (Capraro & Capraro, 2002). The limitation is different: a reliable instrument can still compress a person into a category that discards most of the information.
Why do I recognize myself in a description that also fits my friend?
Because the description was written to be recognizable. The same text given to different people is rated as accurate by most of them, a result established with identical feedback handed out to a whole class (Forer, 1949).
Are the four dimensions stable over time at all?
Rank-order stability is real and rises with age. The ordering of people on a trait is fairly consistent across years while the absolute positions of individuals shift, so the same person can move without the population rearranging itself (Roberts & DelVecchio, 2000).
Should I retake the test to see which result is 'really' me?
Retaking can show you the width of your own range, which is useful. It will not converge on a hidden true score, because part of the variation is the person changing rather than the measurement failing.
Further Reading (primary sources & authoritative references)
- Bess, T. L., & Harvey, R. J. (2002). Bimodal score distributions and the Myers-Briggs Type Indicator: Fact or artifact? Journal of Personality Assessment, 78(1), 176–186. — DOI
- Capraro, R. M., & Capraro, M. M. (2002). Myers-Briggs Type Indicator score reliability across studies: A meta-analytic reliability generalization study. Educational and Psychological Measurement, 62(4), 590–602. — DOI
- Fleeson, W. (2001). Toward a structure- and process-integrated view of personality: Traits as density distributions of states. Journal of Personality and Social Psychology, 80(6), 1011–1027. — DOI
- Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of personality. Psychological Review, 102(2), 246–268. — DOI
- Borkenau, P., Mauer, N., Riemann, R., Spinath, F. M., & Angleitner, A. (2004). Thin slices of behavior as cues of personality and intelligence. Journal of Personality and Social Psychology, 86(4), 599–614. — DOI
- John, O. P., & Robins, R. W. (1993). Determinants of interjudge agreement on personality traits. Journal of Personality, 61(4), 521–551. — DOI
- Forer, B. R. (1949). The fallacy of personal validation: A classroom demonstration of gullibility. Journal of Abnormal and Social Psychology, 44(1), 118–123. — DOI
- Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age. Psychological Bulletin, 126(1), 3–25. — DOI
- Pittenger, D. J. (2005). Cautionary comments regarding the Myers-Briggs Type Indicator. Consulting Psychology Journal: Practice and Research, 57(3), 210–221. — DOI
- Boyle, G. J. (1995). Myers-Briggs Type Indicator (MBTI): Some psychometric limitations. Australian Psychologist, 30(1), 71–74. — DOI
Related reading
- How to type someone else — the same four dimensions, read from outside instead of from inside.
- Cognitive functions, explained — why two people with the same letters are often described as sharing a function order they do not actually share.
- Trait versus type — the difference between a position and a category, in one page.
- Where MBTI stops — the structural problems, stated plainly.
- Your own four positions are on the test page; the letters are only the index to them.