How the Deep Quiz is scored

1. The model: five domains, ten aspects

The quiz measures the ten aspects of the Big Five described by DeYoung, Quilty and Peterson (2007). Each of the five broad domains splits into two empirically distinct aspects, which is what lets the result talk about how your traits combine rather than only where you land. We show plain names; the research names are in brackets.

2. Where the questions come from

The 80 agree-disagree statements are items from the International Personality Item Pool (ipip.ori.org), which places its items in the public domain for any use. We use eight per aspect, four worded towards the high pole and four towards the low pole, so a habit of agreeing with everything cancels out instead of inflating a score. A handful of items were lightly edited for plain English; the original wording is kept in our item bank.

The 10 either-or pairs and 12 situations are our own. Pairs give each domain a second reading in a different format. Situations keep the feel of the animal quiz and give the result concrete moments to point at. Each situation also offers “None of these fits me”, which counts for nothing rather than forcing a bad fit.

Three check items are mixed in: two statements almost nobody can truthfully agree with, and one that asks you to pick a specific answer. They never affect your scores; they only inform the confidence rating.

3. From answers to percentiles

Each statement scores one to five, with low-pole items reversed. Each aspect is the weighted average of its statements, its pairs and its situation choices (situations count at half the weight of a statement). The average is compared with a norm, the typical score and spread for that aspect, to produce a percentile: the share of people who score below you.

At launch the norms are provisional centre points for a young, self-selected online audience. Once a thousand people have completed the Deep Quiz we switch to our own empirical distribution and refresh it monthly. Every saved result records which norms it was scored with.

Percentiles are described in five bands: strongly towards the low pole (under 15), leaning low (15 to 35), balanced (35 to 65), leaning high (65 to 85) and strongly towards the high pole (over 85).

4. The type code

The four familiar letters come from four domains, using the well-replicated links between four-letter type systems and the Big Five: Energy with people gives E or I, Curiosity gives N or S, Warmth gives F or T and Discipline gives J or P, each split at the median. The fifth letter after the hyphen comes from Emotional sensitivity: C for calm below the median, S for sensitive above it.

Every letter carries a lean strength from its distance to the median: slight (under ten points), clear (ten to twenty-five) or strong (more). A slight letter means you could reasonably read as either, and the result says so.

Not affiliated with any trademarked assessment. Type codes here are a summary computed from Big Five aspect scores, not a separate theory of personality.

5. Your animal

Each of the 31 animals has a prototype: a position on all ten aspects, set from the animal’s existing profile. Your animal is the prototype closest to your ten scores (a plain Euclidean distance in standard-deviation units). The match percentage is a display of that distance, scaled so a typical closest match reads around 80 and an exact match reads 100. If your two closest animals are within 0.08 of each other, the headline names both.

“What kind of Fox” lists the aspects where you sit at least three quarters of a standard deviation from the Fox prototype. As completions grow we fit a small per-animal adjustment so no animal becomes rarer than one percent or more common than twelve percent of results.

6. The confidence rating

Results always show, but the quiz is honest about how much to trust them. It checks for disagreement with a statement about reading carefully, agreement with statements almost nobody can truthfully agree with, long runs of identical answers, very fast answering, disagreement between positively and negatively worded items, mostly neutral answers, and many “none fits” picks. No flags means high confidence; one soft flag means medium; two flags or any hard flag means low, with a suggestion to retake on a quieter day.

7. What it is and is not

The Deep Quiz is a reflection tool. It is not a clinical instrument, it does not diagnose anything, and it should never be used to select or screen people for jobs, teams or anything else. Both ends of every scale are described as a way of operating, never as a deficit.

It is free for individuals and supported by ads, including four short screens between the parts. We publish our reliability and agreement statistics on this page once enough people have taken it.