Methodology
How these tests are built
Most personality sites will not tell you how their test works. Here is ours, including the parts that are not flattering.
What each test is based on
Built on the two-dimension model of adult attachment — attachment anxiety and attachment avoidance — that underlies the Experiences in Close Relationships research tradition. The nine styles are positions on those two continuous dials, not nine separate boxes. The items are our own wording, written to match that model; they are not the original research scale.
- Length
- 30
- Item format
- 18 forced-choice scenarios · 12 agree/disagree statements
- Scored dimensions
- Relationship anxiety · Closeness avoidance
- Result space
- 9 types, assigned by position on both dials
- Reverse-scored items
- 5 of 12 statements
Built on the money-script framework: four learned beliefs about money — avoidance, worship, status and vigilance — identified in Klontz's work on financial psychology. Most people run two scripts at once, which is why the test can return a blend rather than a single script.
- Length
- 30
- Item format
- 18 forced-choice scenarios · 12 agree/disagree statements
- Scored dimensions
- Money avoidance · Money worship · Money status · Money vigilance
- Result space
- 10 types, assigned by highest-scoring dimension, with blends
- Reverse-scored items
- 4 of 12 statements
Built on five dimensions that recur across observational couples research: how you handle conflict, how much togetherness you need, how you express affection, how much you want planned, and how readily you say what is going on inside. This is the least standardised of the three tests, and it is the one we would defend least strongly as measurement.
- Length
- 30
- Item format
- 20 forced-choice scenarios · 10 agree/disagree statements
- Scored dimensions
- Conflict style · Togetherness · Expression · Planning · Emotional openness
- Result space
- 9 types, assigned by position on both dials
- Reverse-scored items
- 5 of 10 statements
How scoring works
Every answer adds points to one or more dimensions. Your total on a dimension is divided by the highest total that dimension could reach, which gives a number from 0 to 100. That number is your position on the scale — it is not a percentile, and it does not mean you scored higher than that percentage of people.
Most of the questions are forced-choice scenarios rather than agree/disagree statements. That is deliberate. Agree/disagree items are vulnerable to yes-saying: some people agree with everything, which quietly inflates every dimension at once. A scenario makes you choose, so agreeing with everything is not an available move.
For the same reason, every dimension includes at least one reverse-scored statement, where agreeing lowers the score instead of raising it. If someone simply agrees their way down the page, those items pull against the rest and the profile flattens out instead of pretending to be a strong result.
Where a test sorts you into a type by cutting a dial into low, moderate and high, the cut-offs are our editorial choice, not a statistical norm. We do not yet have enough takers to publish real percentiles. When we do, this page will say so, and the figure will appear on your result page. We would rather leave that blank than invent it.
How the paid report is written
Your free result is computed from your scores. The paid report is not.
The moment you pay, your full answer sheet — all thirty of your specific choices, in the words you actually picked, along with your dimension scores — is sent to Claude, which writes the report for you, then and there. It quotes your answers back to you because it has them. It notices where two of your answers contradict each other, because it can see both.
This is the real difference from the large personality sites. Their paid reports are one document per type, written once and sold to everyone who lands on that type — which is why their most common complaint is that the paid version reads like a longer edition of the free one. Yours did not exist before you bought it, and nobody else has a copy.
The honest trade-off: a report written on demand is not read by a human before you see it. That is why we publish a complete sample report, generated the same way, so you can judge the writing before you spend anything. We would rather you read the sample and decide against it than buy something that disappoints you.
What this is not
Not a clinical assessment, and not a diagnosis. Nothing here identifies a disorder or replaces an assessment by a qualified professional.
Not a validated instrument. The frameworks these tests are built on are researched and well documented. Our specific thirty-item adaptations of them are not: we have not published reliability coefficients or run a validity study on these exact items, and we are not going to imply otherwise by putting a number next to the word accuracy.
Not a prediction. Attachment patterns shift, money beliefs shift, and couples change how they operate. A result describes how you answered this week, which is useful, and less permanent than a type label makes it sound.
Not the right tool for something heavy. If a result opens something painful, a therapist trained in that area will get you further than any quiz. That is not a disclaimer; it is genuinely the faster route.
Questions about any of this: hello@knowmyself.co.
Last updated: September 2026