Methodology

Personaspect measures the five traits of the Big Five model with a public-domain questionnaire, scores it against published norms and turns the scores into a result you can read in a minute. This page sets out each step, with its sources and its weak points.

The questionnaire

The statements come from the IPIP-NEO-120, a 120-item inventory built by John A. Johnson from the International Personality Item Pool, which is in the public domain (Johnson, 2014). It measures 30 facets, six per trait, with four statements each. You answer each statement on a five-point scale from "very inaccurate" to "very accurate".

The test starts with a first block of 60 statements, two per facet. Fifty-two of them are the statements the IPIP-NEO-120 shares with the IPIP-NEO-60, a validated short form (Maples-Keller et al., 2019). For the eight facets where the two forms share only one statement, we added the other IPIP-NEO-120 statement of that facet with the highest item-total correlation in Johnson's data. The result is a subset of the full questionnaire, so the same trait norms can be applied, but they were computed on 24-statement traits: after 60 statements the percentile is an approximation, and the 60-statement block has not been validated as a form of its own. After your result you can answer the remaining 60 statements and get the full IPIP-NEO-120 without starting again.

The Italian version

No validated Italian translation of the IPIP-NEO-120 has been published. We started from the Italian translation by Michelangelo Vianello published on the IPIP site, kept its wording unless a statement added or dropped meaning or only fitted one gender, then had every statement translated back into English blind, without the originals, and compared the result with the English items. In all, 38 of the 120 statements now differ from Vianello's wording. Every statement now reads correctly for anyone, because the test never asks your gender. This is a documented translation, not a validated one; a native-speaker review is still pending, and we treat it as a known risk.

Scores and norms

Each trait score is the average of your answers to its statements, with reverse-worded statements flipped. To say where a score stands among other people, we compare it with the norms of Kajonius and Johnson, who published means and standard deviations for every trait and facet of the IPIP-NEO-120 from 320,128 online respondents in the United States (Kajonius and Johnson, 2019). From those we compute a percentile, assuming the scores follow a normal distribution.

Three limits follow. The sample is American, young (average age 28) and self-selected online, and we apply it to Italian users too, because no Italian norms exist. The authors themselves note that three facets (emotionality, intellect and altruism) are probably inflated in their sample. And facet norms describe four-statement facets, so after the 60-statement block facets get no percentile at all.

What the free result shows

The free result places each trait on one of seven steps, from clearly low to clearly high, derived from the percentile. It shows no numbers on purpose: a percentile looks more precise than a self-report questionnaire is, and the steps keep the result honest about that. The words for each step describe a direction ("more practical than curious"), not a verdict.

How the archetypes are built

The 16 archetypes are our own synthesis, not a finding of personality research. For Openness, Conscientiousness, Extraversion and Agreeableness, a percentile of 50 or above counts as high and below 50 as low, which gives 2 × 2 × 2 × 2 = 16 combinations. Neuroticism does not change the archetype; it sets the calm variant (below 50) or the intense variant (50 or above).

A median split throws information away. Someone at the 49th percentile and someone at the 51st get different letters although their scores are almost the same, and the archetype says nothing about how far from the middle you are. In the norm sample each trait splits half and half by construction; among our users it need not, and since traits are correlated the 16 archetypes are not equally common. We will publish real frequencies once there is enough data. Read the archetype as a memorable name for your pattern, and the trait positions as the result.

Answer quality checks

The result flags three patterns that make answers less trustworthy: a median time below two seconds per statement, a threshold drawn from research on careless responding (Huang et al., 2012); ten or more identical answers in a row, which are rare in honest IPIP data (Johnson, 2005); and opposite extreme answers within the same facet (for example fully agreeing with a statement and with its opposite) in five or more facets. A flag adds a visible note to the result. It never changes a score. The thresholds are provisional and will be checked against real data.

How the pages are written

The pages of this site are written by the Personaspect editorial team, with AI assistance for drafting, from the sources they cite. Every study cited is checked against its publication record, and numbers come only from those sources. No page is written or reviewed by a clinician, and nothing on the site is medical or psychological advice. If you find an error, write to us and we will correct it.