How this test is built
The frameworks, the scoring, and the parts we don't claim
What the test actually measures
Most tests in this space measure one thing. Attachment quizzes tell you you're anxious and stop there. Love-language quizzes tell you it's acts of service and stop there. Neither is much use on a Tuesday night when you're arguing about something that isn't the thing you're arguing about.
So this one measures four axes and combines them. Five questions each, twenty in total.
| Axis | Questions | Framework it draws on | What it produces |
|---|---|---|---|
| Attachment | 1–5 | Attachment theory, in the two-dimensional anxiety × avoidance form used by the ECR family of self-report scales | Two scores, 0–100. Their combination gives one of four quadrants: secure, anxious, avoidant, disorganized |
| Conflict style | 6–10 | The Thomas-Kilmann five modes | A dominant mode: competing, collaborating, compromising, avoiding, accommodating |
| Love language | 11–15 | Gary Chapman's five love languages | A primary and a secondary language, out of words, quality time, gifts, acts of service, touch |
| Secure behaviours | 16–20 | Our own block, drawn from what the attachment literature describes as security-supporting behaviour | One score, 0–100 — the habits axis |
The first three axes come from published work that we've adapted rather than reproduced. The fourth is ours, and we say so.
How 20 answers become one of 12 types
Every answer option carries a fixed weight. Choosing "something's wrong, did I do something?" when a text goes unanswered for six hours adds three points to the anxiety counter; "honestly didn't notice" adds two to avoidance instead. Nothing is interpreted, weighted by a model, or adjusted after the fact. The same 20 answers always produce the same result.
Step one — the quadrant
Anxiety and avoidance are each summed, then normalised to 0–100. Fifty is the cut on both:
- Low anxiety, low avoidance → secure
- High anxiety, low avoidance → anxious
- Low anxiety, high avoidance → avoidant
- High anxiety, high avoidance → disorganized, also called fearful-avoidant
Step two — the column
Your primary love language sorts you into one of three groups: words and quality time land in connection, acts of service and gifts in provider, physical touch in its own column. If two languages tie, the secondary breaks it.
Step three — the type
Four rows by three columns is twelve cells, and each cell is a named type. Anxious plus touch is The Firework. Avoidant plus acts of service is The Fortress. Secure plus words is The Lighthouse. The full grid is here.
Your conflict style and your secure-behaviours score don't change which type you get. They fill it in — conflict style drives the "how this type argues" section, and the secure-behaviours score calibrates how strongly the pattern is currently running. A high score means the tendency shows up as a tilt; a low one means it's doing most of the driving.
Why the love language picks the column and the conflict style doesn't. Love language varies independently of attachment, so it spreads people evenly across the three cells of a row. Conflict style doesn't — nearly every avoidant person scores "avoiding", which would collapse the whole row into one type. The grid is built the way it is to preserve the distribution.
Where the percentages come from
Right now they're seeds, and every page that shows one says so.
Published research puts secure attachment at roughly half the population, anxious and avoidant at around a fifth each, and disorganized at somewhere near one in twenty. We took those proportions, split each row across the three love-language columns using the relative frequency with which each language is reported as primary, and used the result as a starting estimate for all twelve types. They sum to 100%.
That is a reasonable place to start and a poor place to finish. It's an estimate built from other people's numbers, so no page presents it as a measurement of our own test-takers, and no page rounds it into something that looks more precise than it is.
Once enough tests are complete, a counter replaces the seeds with the real distribution, and every stat gets a sample size and a date next to it. Two rules hold from day one:
- We never publish a number without saying what it's a number of.
- We never publish a segment — by region, age or anything else — below 500 completions. Below that, the global figure is the honest answer.
One thing worth stating plainly even after the counter goes live: people who take an online love quiz are not a random sample of humanity. Our distribution will describe our test-takers, and that is all it will ever describe.
What we don't claim
- This is not a validated psychometric instrument. It hasn't been through reliability or validity testing, and we're not going to imply it has by putting the word "scientific" near it.
- It is not a diagnosis of anything. No disorders, no clinical language, no verdicts. Patterns and tendencies only.
- It is not therapy and doesn't replace it. If something here lands close to something painful, a licensed professional will take you further than a quiz will.
- The frameworks themselves are contested. The five love languages in particular are popular rather than empirically established as a typology. We use them because they're a genuinely useful vocabulary for talking to a partner, not because they're settled science.
- It's for adults. The test is written for people 18 and over.
What happens to your answers
Nothing leaves your browser. Scoring runs entirely in the page — the twenty answers are turned into a type on your own device, and no answer is ever transmitted to us. There's no email wall, no account, and no way for us to look up what you said, because we never had it.
The only thing that travels is an anonymous increment to a completion counter, which is what will eventually turn the seeded percentages into real ones. It carries no answers and nothing that identifies you. The privacy policy has the full detail.
How this gets updated
Type pages are static and rebuilt when their content changes; every one carries a last-updated date. The statistics change on their own schedule, when the counter has enough behind it to justify replacing a seed. If a number on this site changes, it's because the data changed — not because a better-looking figure was available.
Frequently asked questions
Is the Love Blueprint test scientifically validated?
No, and we won't claim otherwise. It's inspired by published frameworks — attachment theory, the five love languages, the Thomas-Kilmann conflict modes — but it has not been through psychometric validation. Treat it as a well-built mirror, not an instrument.
Is this a clinical diagnosis?
No. Love Blueprint is a self-discovery and entertainment tool based on popular relationship frameworks. It is not a psychological assessment or medical advice.
Where do the rarity percentages come from?
They are starting estimates, seeded from published attachment-style prevalence research and split across the three love-language buckets. They are not yet a count of our test-takers. Once enough tests are complete, a live counter replaces the seeds and every page states the sample size and date.
Do you store my answers?
No. The test is scored entirely in your browser. Your answers are never sent to a server, and you don't need an email address to get your result.
Can my type change?
Yes. Attachment patterns are learned expectations about closeness, and they move with experience. Retaking the test after a significant relationship change is the intended use, not a sign the first result was wrong.