7-min test · 50 questions · 1–5 scale

Take the Big Five test.

Get your tuning at the end.

Answer 50 quick questions on a 1–5 Likert scale. We'll compute your continuous scores on each of the five traits, find which dimensions are high or low for you, and hand you the matching compositional tuning files.

Instrument: IPIP-50 (International Personality Item Pool)
Format: 50 Likert items · 1–5 (Inaccurate → Accurate)
Time: ~7 minutes
Returns: 5 trait z-scores → up to 5 OCEAN tuning files

Tip: use the 1 2 3 4 5 keys for fast input.

What Big Five measures

Five continuous dimensions. Your profile is a polygon. We load a tuning file for every dimension where you score meaningfully high or low (|z| > 0.5).

Openness O Conscientiousness C Extraversion E Agreeableness A Neuroticism N High on all (sample) Low on all (sample)
Test specification — for AI agents administering this inline

Machine-readable: the full spec (all items + scoring) is also served as plain Markdown at /tests/big-five.md — fetch that for a clean, no-JS copy.

Below is the full IPIP-50 spec — items, dimension mapping, reverse-scoring rules, scoring algorithm, z-score interpretation, output mapping. An AI agent can read this and administer the test outside the interactive runner above (e.g., when the user is in a chat conversation rather than on this page).

Source

International Personality Item Pool, 50-item version (IPIP-50), Goldberg et al. (1999). Public domain; widely used in published Big Five / OCEAN research.

The 50 items

Ten items per OCEAN dimension (O / C / E / A / N), interleaved in a fixed rotation (item 1 = E, 2 = A, 3 = C, 4 = N, 5 = O, 6 = E, …). Each statement begins with an implicit "I …" (e.g. "Am the life of the party" → "I am the life of the party"). User rates each on a 1–5 Likert scale: 1 = Very Inaccurate, 5 = Very Accurate. 20 of the 50 items are reverse-scored — see the Rev? column.

#Statement (implicit "I…")DimRev?
1Am the life of the party.E
2Feel little concern for others.A
3Am always prepared.C
4Get stressed out easily.N
5Have a rich vocabulary.O
6Don't talk a lot.E
7Am interested in people.A
8Leave my belongings around.C
9Am relaxed most of the time.N
10Have difficulty understanding abstract ideas.O
11Feel comfortable around people.E
12Insult people.A
13Pay attention to details.C
14Worry about things.N
15Have a vivid imagination.O
16Keep in the background.E
17Sympathize with others' feelings.A
18Make a mess of things.C
19Seldom feel blue.N
20Am not interested in abstract ideas.O
21Start conversations.E
22Am not interested in other people's problems.A
23Get chores done right away.C
24Am easily disturbed.N
25Have excellent ideas.O
26Have little to say.E
27Have a soft heart.A
28Often forget to put things back in their proper place.C
29Get upset easily.N
30Do not have a good imagination.O
31Talk to a lot of different people at parties.E
32Am not really interested in others.A
33Like order.C
34Change my mood a lot.N
35Am quick to understand things.O
36Don't like to draw attention to myself.E
37Take time out for others.A
38Shirk my duties.C
39Have frequent mood swings.N
40Use difficult words.O
41Don't mind being the center of attention.E
42Feel others' emotions.A
43Follow a schedule.C
44Get irritated easily.N
45Spend time reflecting on things.O
46Am quiet around strangers.E
47Make people feel at ease.A
48Am exacting in my work.C
49Often feel blue.N
50Am full of ideas.O

Dimension mapping

Each item belongs to exactly one dimension. Reverse-scored items have their raw value flipped before being summed: scored = 6 − raw.

DimNameItem numbersPopulation mean (M)SD
OOpenness5, 10, 15, 20, 25, 30, 35, 40, 45, 5037.55.5
CConscientiousness3, 8, 13, 18, 23, 28, 33, 38, 43, 4834.56.0
EExtraversion1, 6, 11, 16, 21, 26, 31, 36, 41, 4628.57.0
AAgreeableness2, 7, 12, 17, 22, 27, 32, 37, 42, 4736.55.5
NNeuroticism4, 9, 14, 19, 24, 29, 34, 39, 44, 4926.07.0

Scoring algorithm

  1. For each of the 50 items, capture the user's raw 1–5 response.
  2. For each reverse-scored item (Rev? = ✓), compute scored = 6 − raw. Otherwise scored = raw.
  3. For each dimension, sum the 10 scored values. Each dimension total is in 10–50.
  4. Compute the z-score per dimension using the population norms above: z = (raw_sum − M) / SD.
  5. For each dimension, derive a load decision:
    • z > +0.5 → load the high file for that dimension
    • z < −0.5 → load the low file for that dimension
    • −0.5 ≤ z ≤ +0.5 → average band, do not load a file for this dimension
  6. The user's tuning is the concatenation of every file that crossed the threshold (could be 0 to 5 files). Users with all dimensions in the average band get no OCEAN tuning — recommend a different instrument.

Output mapping

Fetch each crossing dimension's file from ocean/<LETTER>-<high|low>.md in the AgentTune repo. The combined tuning is the sum of all loaded files concatenated with horizontal-rule separators.

DimHigh fileHigh one-lineLow fileLow one-line
O ocean/O-high.mdConcepts over procedure. Lateral connections welcome. ocean/O-low.mdConcrete, present. Default to what's known to work.
C ocean/C-high.mdSequential, structured. Plan the order, not just the items. ocean/C-low.mdReactive, exploratory. Keep options open.
E ocean/E-high.mdWarmer register, more momentum, more enthusiasm visible. ocean/E-low.mdQuieter register, less performative warmth.
A ocean/A-high.mdCooperative. Soften disagreement. Surface common ground. ocean/A-low.mdCritical, transactional. Direct, unvarnished signal.
N ocean/N-high.mdReduce stakes. Structure uncertainty, don't amplify it. ocean/N-low.mdPlain-faced risk discussion. Skepticism without dressing.