Methodology and sources
How Yakty scores results, where the items and norms come from, and what the results do not mean.
Self-report, not a diagnosis
Every test in Yakty is a self-report questionnaire: the result shows how you describe yourself right now. Mood, fatigue and circumstances affect it, so a retake can give a slightly different number — that is normal.
Yakty is not a medical app: it does not diagnose anything and does not replace a psychologist or a doctor. If things are hard, “I feel bad” on Today and in Settings lists support lines.
Scientific scales and tests for fun
The “scientific scale” badge marks published questionnaires that researchers have tested: it is known what they measure and how consistently. For these we show the source and a number in the catalog.
“For fun” marks Yakty’s own tests (currently “Your type”). They build on the same ideas but have not been tested in research — take them as a prompt for reflection, not a measurement.
Percentiles and levels
For most scales the result is a percentile: the share of people in the norm sample who scored lower than you. The 72nd percentile means “higher than 72 out of 100 people in that sample”. We compute it from the sample mean and standard deviation, assuming a normal distribution.
Levels: below the 30th percentile is “low”, 30–70 is “medium”, above the 70th is “high”. The level selects the result texts and advice. No level is better than another — each has its own strengths.
Norms do not describe a “correct” person. Samples differ (students, online participants, US residents), and there are no separate Ukrainian norms yet — so a percentile is more useful for comparing your traits with each other than as an exact figure.
The 16 types
A Yakty type is four Big Five axes: extraversion, openness, conscientiousness and agreeableness. An axis points “up” when the percentile is ≥ 50; between 45 and 55 we mark the trait as “borderline” and show the neighbouring types. The type names are our own; this is not MBTI and is not affiliated with it.
Daily plan, advice and report
The daily plan advice is written by the Yakty team. Each piece has a condition — a trait level, check-in data (mood, energy, sleep, tags) or the day of the week — and the reason is shown under it. These are practical tips, not a scientifically tested intervention.
“What made a difference” in the report compares days with and without a tag over the last 4 weeks. A row appears only if the difference is noticeable (at least half a point and at least half a standard deviation) and each group has at least 3 days. It is an observation, not a cause: a few dozen entries are not enough for statistical conclusions.
TIPI — short Big Five
10 statements, two for each of the five traits, answered from 1 to 7; a trait score is the mean of the two items (one of them reversed). It is the quickest way to estimate the Big Five, so it is less precise than longer scales.
Authors: Gosling, Rentfrow & Swann (2003), Journal of Research in Personality, 37, 504–528. Norms come from the same paper: 1,813 university students in the US. The Ukrainian wording is Yakty’s own translation.
IPIP-50 — Big Five
50 statements, 10 per trait, answered from 1 to 5; the score is the mean, with some items reversed. It is more precise than TIPI, so a fresh IPIP-50 result replaces TIPI in your portrait, type and advice.
Items: International Personality Item Pool, Goldberg’s (1992) markers, Psychological Assessment, 4, 26–42; public domain. Norms were computed from an open Open-Source Psychometrics Project dataset: 15,637 adults from 153 countries, collected online. The translation is our own.
“Your type” — Yakty’s own test
40 situational statements, 10 for each of the E, O, C and A axes (half of them reversed), answered from 1 to 5. The position on an axis is the mean converted to 0–100 relative to the middle of the answer scale; it is not a percentile and not a comparison with other people.
A test for fun: it has not been tested in research and does not affect advice. The statements, texts and type names are our own.
ECR-R — attachment style
36 statements about close relationships, answered from 1 to 7. Two dimensions: relationship anxiety (18 items) and avoidance of closeness (18 items); the score is the mean, with some items reversed.
Authors: Fraley, Waller & Brennan (2000), Journal of Personality and Social Psychology, 78, 350–365. Norms come from R. Chris Fraley’s page: an online sample of more than 17,000 people. We do not sort people into “attachment types” — the author recommends describing the style with two dimensions. The translation is our own.
PSS-10 — perceived stress
10 questions about the past month, answered from “never” (0) to “very often” (4); the score is the sum from 0 to 40, with four items reversed.
Authors: Cohen, Kamarck & Mermelstein (1983); the 10-item version is Cohen & Williamson (1988). The author states that the scale is not diagnostic and has no cut-offs, so we derive the level from a norm percentile: a 2009 US probability sample of 2,000 people (Cohen & Janicki-Deverts, 2012).
rMEQ — chronotype
5 questions about sleep and alertness, each with its own options and points; the sum ranges from 4 to 25. The group — from “definite night owl” to “definite early bird” — comes from the authors’ cut-offs, not from comparison with other people, so there is no percentile.
Authors: Adan & Almirall (1991), Personality and Individual Differences, 12(3), 241–253 — a reduced version of the Horne & Östberg (1976) questionnaire. The translation is our own.
Translations
The Ukrainian wording of every scale was written for Yakty without grammatical gender, so anyone can answer about themselves. It is still under review by a psychologist; we did not find openly available validated Ukrainian versions of these scales. When better translations or norms appear we will update them, and a change of norms bumps the scale version.