How we score your test

The questions

Both tests use inventories from the International Personality Item Pool (IPIP), an open bank of personality items that researchers created so the Big Five could be measured without licensed commercial tests. IPIP scales are used in hundreds of peer-reviewed studies.

• Quick test: the 50-item IPIP Big-Five Factor Markers (Goldberg, 1992), 10 statements per trait.
• In-depth test: the IPIP-NEO-120 (Johnson, 2014), 120 statements measuring the five traits and 30 facets, modeled on the facets of the commercial NEO-PI-R.

You rate each statement from 1 (very inaccurate) to 5 (very accurate). Some statements are worded in the opposite direction and are reverse-scored. Your raw score for a trait is the sum of its items: 10 to 50 on the quick test.

How your percentiles are calculated

A raw score only means something next to other people's scores, so each one is converted to a percentile. We compare you with 602,011 people who answered the same 50 questions, from the public dataset released by the Open-Source Psychometrics Project (collected online in 2016–2018).

Before calculating anything we removed responses that weren't usable: 140,907 with unanswered items, 271,112 repeat submissions from the same network address, 952 finished in under a minute, and 359 that gave the same answer to every question.

We checked what that filter does. Adding the 269,666 repeat-address responses that pass our other checks to the 602,011 we use moves a percentile by at most 2.7 points anywhere in the table. As a group they differ from everyone else by no more than 0.21 standard deviations on any trait (the largest is openness, lower). That group mixes people on shared networks and VPNs with people who took the test more than once, so it can't show how any one of those subgroups scores.

Your percentile is the share of that sample who scored lower than you, plus half of those who scored exactly the same. The 50th percentile is the middle. Results are shown between 1 and 99, since no single test can place anyone at the absolute extremes.

The in-depth test

The in-depth test has norms of its own, from John A. Johnson's IPIP-NEO data repository: 410,376 people who answered all 120 items of the same inventory (208,774 responses with unanswered items were left out). This is the sample the IPIP-NEO-120 was developed on. Both the five domain scores and all 30 facet scores are reported as percentiles against it.

Quality checks

In the norm sample, every one of the 50 items correlates positively with the rest of its trait's items (from 0.31 to 0.72). That confirms each question is scored in the right direction and measures the same thing as the others on its scale.

For the in-depth test, each facet's four items hang together well (coefficient alpha from 0.64 to 0.88, average 0.76), which is typical for four-item scales, and facets correlate more strongly with the other facets of their own trait than with other traits.

What the numbers can't tell you

• The comparison group is online test-takers. They volunteered on a free website, skew young, and don't represent any one country. There are no separate norms by age or sex.
• It's self-report. Scores reflect how you see yourself, and a stressful or low period can shift your answers. Why scores shift
• Small differences aren't meaningful. Every questionnaire has measurement error, so the gap between the 55th and 62nd percentile could reverse on a retake. How to read your percentiles
• It isn't a diagnosis. Personality scores are for self-reflection, not for clinical or hiring decisions.
• Paid reports are written by an AI model from your scores. The scores themselves are calculated as described here and never by AI.

Sources

• Goldberg, L. R. (1992). The development of markers for the Big-Five factor structure. Psychological Assessment, 4(1), 26–42.
• Johnson, J. A. (2014). Measuring thirty facets of the Five Factor Model with a 120-item public domain inventory: Development of the IPIP-NEO-120. Journal of Research in Personality, 51, 78–89.
• International Personality Item Pool: ipip.ori.org
• Open-Source Psychometrics Project, IPIP Big Five Factor Markers response data (IPIP-FFM-data-8Nov2018): openpsychometrics.org/_rawdata
• Johnson's IPIP-NEO data repository: osf.io/tbmh5