Why Your MBTI Results Change Every Time: Check Your Near-50/50 Axis First
Does your free MBTI test give a different result every time? 11 of 15 saved SCHROE results had an axis at 50–55%. A table for reading your % and tips for retaking.
SCHROE Editors
Friday lunch at the noodle place across from the office. Mia slides her phone to the middle of the table. “It changed again. Last month I was ISTJ. This morning it says ISTP.” Dana, sitting across from her, took the same test that morning too. “Weird. I get ENFP every single time.” They put their result screenshots side by side and compare the bars. All four of Dana’s bars stretch out toward the ends. One of Mia’s bars barely leaves the center line.
At 11 that night, after working late, Mia opens the test again on the subway. At “If a friend suddenly says they're nearby and asks you to come out after work?” her thumb goes straight to the far end of “I want to rest tonight and suggest another time.” At “When you pause at night and think about tomorrow?” she picks “I leave tomorrow's plans for tomorrow.” On a slow Sunday morning after sleeping in, her thumb would probably have stopped somewhere near the middle, at 50/50.
Mia and Dana are made up to make the point easier to see. But the difference between their two screens, a bar stuck to the center versus a bar that runs all the way out, shows up just as clearly in real results saved on the SCHROE site. So let’s look at why your MBTI results change from one free test to the next, by counting those saved results one by one.
The number on your result screen tells you how far you lean
First, a quick guide to the result screen. The SCHROE MBTI-style test shows four axes as bars: Introversion–Extraversion, Sensing–Intuition, Thinking–Feeling, and Judging–Perceiving. Each bar starts at the center and grows toward the side you lean to, with a label underneath such as “Judging (J, 54%).”
That number is always 50 or higher, because it shows only the share of the side you lean toward. 50% means the two sides are exactly even; 100% means every answer on that axis went to one end. When the two sides score exactly the same, you see “Evenly matched (50% / 50%)” instead of a number.
So 54% doesn’t mean “54 points of Judging.” It means “leaning very slightly toward Judging.” An 85%, on the other hand, means one side came through quite clearly on that axis. The closer the number is to the middle, the more room there is for the letter to flip next time. We’ll get to why in a moment.
We counted all 60 axes in 15 saved results
As of October 3, 2026, the SCHROE site has 15 saved results for the human MBTI-style test once we leave out a test account used to check the site. With four axes per result, that makes 60 numbers. We recalculated each one exactly the way the site shows it on the result screen, then counted the axes that landed between 50% and 55%.
- 13 of the 60 axes showed 50–55% on the result screen. None was an exact tie; all 13 were between 51% and 55%.
- 11 of the 15 results had at least one such axis. Two had two.
- Only 4 results had all four axes at 56% or more.
One thing to be clear about before reading these numbers: 15 is a very small sample. It only includes results that someone saved after logging in. We left out 17 results saved by a test account used to check the site, since they aren’t real users’ results; the remaining 15 each come from a different account, saved once. We also have no records of the same person retaking the test and comparing results, so this article does not say “X% of people got a different result.” The only point here is not a rate. It’s this: an axis sitting near 50/50 is not rare.
Split by axis, Judging–Perceiving came up most often with 6 results, Introversion–Extraversion and Thinking–Feeling had 3 each, and Sensing–Intuition had 1. Numbers this small can’t tell us that Judging–Perceiving is naturally shakier, though. Take it only as a sign that any axis can end up near 50%.
A result stuck to the center, and one that runs all the way out
Here are two of the saved results side by side. Names are hidden; only the numbers from the result screen are shown.
Result A is ENTP. Extraversion 68% and Intuition 63% lean a fair way, but Thinking 51% and Perceiving 51% sit almost on the center line. Result B is ENFP. Extraversion 82%, Intuition 78%, Feeling 85%, and Perceiving 81%: all four bars stretch out toward the ends.
Back at the noodle place, Result A is Mia’s screen and Result B is Dana’s. Same test, same 40-question format, yet one could easily show a different letter next time while the other will most likely stay put. What makes the difference isn’t whether a personality is “strong” or “faint.” It’s how evenly both sides showed up on that axis.
At 51%, one step on one question can flip the letter
Why do axes near the middle flip so easily? It makes sense once you look at how you answer.
Every question has a seven-step scale. One end is “Strongly left,” the other is “Strongly right,” and the middle is “50/50.” The closer you go to an end, the more points that side gets, and the 50/50 middle adds nothing to either side. In one run you only get 10 questions per axis, and each question carries a slightly different weight. Working through the site’s own scoring, moving one answer by one step shifts the result-screen number by roughly 1 to 3.
A number near the middle can come about in two ways. You picked 50/50 a lot, or your answers split, some going to one end and some to the other. Either way, it looks the same on the result screen: somewhere around 50%.
So an axis at 51% is close enough that moving a single answer one step the other way could flip the letter. Even at 55%, you’re only a few steps away. To flip Result B’s Feeling 85%, more than ten steps would have to go the other way. That’s why a changing result doesn’t mean “the test is broken” or “you have no personality.” Your scale was simply close to level to begin with.
The reasons the same person answers differently boil down to about three.
The questions change every time. Each run draws 10 questions per axis from a pool of 100, so 40 in all, at random. We explain how this works in detail in our Dog MBTI test guide.
Your mood and the moment you picture change. “If a friend suddenly says they're nearby and asks you to come out after work?” lands differently after a late night than on a well-rested Sunday. In a deadline week, “I mentally organize tomorrow's to-do list before sleeping” might feel closer; on vacation, “I leave tomorrow's plans for tomorrow” might. Both are genuinely you.
Plenty of people simply have both sides. An axis near 50% means both sides showed up about equally on that axis. For people like that, answers shifting with the situation is actually the natural thing.
When it’s an exact tie, the site now leaves the letter open
Today, when both sides of an axis score exactly the same, the site doesn’t assign a letter for it. Instead of a four-letter type, you see a screen titled “Some preferences are evenly matched,” with a note to think about your usual behavior and review your answers, plus a “Take the assessment again” button. Saying that axis is still open felt more honest than forcing a letter.
Free test vs. official test: why results can differ
You might be thinking, “Doesn’t it change just because it’s a free test?” Let’s be clear first: the SCHROE MBTI-style test is a reflection tool built on SCHROE’s own questions. It is not the official MBTI assessment.
An official test may differ from ours in its questions, in how it scores answers, and in how it reports results. So different results from the two aren’t strange, and neither one proves the other wrong. If one of your axes sat near 50% on our test, seeing a different letter on another test is entirely possible. This article doesn’t go into the details of the official test; it sticks to why ours can wobble.
Find your axis % in this table
Open your result screen and match each of your four numbers to this table. A “step” means moving one answer one notch on the seven-step scale. Step counts are rough estimates based on the site’s scoring.
| Result screen % | How to read it | Distance to a flipped letter (rough) | What to try | |---|---|---|---| | Evenly matched (50%) | Both sides scored the same, so no letter was assigned | None | Read both type descriptions and retake on another day | | 51–55% | Nearly 50/50 | 1 to 5 steps | Read both type descriptions side by side | | 56–60% | A slight lean | 3 to 10 steps | Read your type first, then skim the type one letter away | | 61–70% | A clear lean | 4 to nearly 20 steps | Focus on your own type’s description | | 71% or more | One side stands out | 8 steps or more | An axis that will usually come out the same |
If one of your axes is near 50%, try this
Read both type descriptions side by side. One saved result was ISTJ with Judging at 52%. In a case like that, open both ISTJ and ISTP and see which scenes feel more familiar. On your result screen, expand “How This Type Runs” and you’ll find an “Is This Really Your Type?” checklist. If a lot of you shows up under “It may not fit if...,” the type one letter away might fit better.
Match the conditions when you retake. If you took it on a weekday evening after dinner last time, try a similar time again, on a day when you’re not worn out. Answer with what you actually did last week in mind, not “this is just how I am.” Save the 50/50 notch for when it truly is 50/50; if one side is even a little closer, move one step toward it.
Line up two or three results and look at which letter changed. If it’s always the same axis, that’s your 50/50 axis. It also means the other three letters are a fairly solid part of you. In that case, writing it down as “IST, with Judging and Perceiving about even” describes you more accurately than locking in four letters.
Next Friday, Mia might show up with a result from a Sunday morning. Whether the letter flipped again or not, she now knows to check which bar is standing in the middle first. Curious about your own bars? Open the free MBTI test and start with the numbers on your four axes. Once you have a result, all 16 types at a glance makes it easy to look up the neighboring type too.
MBTI results are for fun and self-reflection. Please don’t use them as a diagnosis or an evaluation. All figures are counts from 15 human MBTI-style results saved on schroe.kr as of October 3, 2026 (test account excluded), recalculated the same way the site’s result screen shows them.