How the questionnaires work

ProfileClaw currently offers seven original, exploratory self-report questionnaires. The published models behind some category names informed the design. Their validity evidence does not transfer to our items. The revised 2026-09.1 questions are editorial candidates: we have not established participant understanding, reliability, validity, representative norms, or equivalence across languages.

What each answer means

| Questionnaire | Purpose of its five-point scale | Items | Areas | | ------------------------------ | ---------------------------------------------------------- | ----: | ----: | | RIASEC interests | Appeal of trying a described activity | 18 | 6 | | Big Five work-style reflection | How the statement usually describes you over recent months | 15 | 5 | | Work values | Importance of a work condition now | 12 | 6 | | Skill confidence | Confidence about doing a specific task | 10 | 5 | | Color-inspired preferences | Usual preference; colors are labels, not identities | 12 | 4 | | DISC-inspired preferences | Usual preference; not a licensed DISC instrument | 12 | 4 | | Work approach preferences | Usual preference; not a situational judgement test | 12 | 4 |

The five response labels remain on one meaning within a questionnaire. There is no default answer or time limit. Choose Cannot judge if you have no basis for rating a prompt. This response remains unscored: it is never turned into a low, neutral or high rating. A true SJT would need concrete scenarios, action choices and a reviewed scoring key; the current work-approach questionnaire has none. The situational-judgement URL and instrument ID remain for existing links and saved records, but do not describe the revised question type.

Scores, versions and incomplete responses

mean-v2 averages only rated items in each area and rounds to two decimal places. An area needs at least two rated items to receive a mean. With zero or one rating, it has no score (null) rather than 0 or 1. The report shows how many items were rated in each area, including areas without a mean. If any answer is Cannot judge, the result presents available arithmetic scores and method limits, but withholds personal pattern interpretation. This avoids inferring a type from uneven evidence. A displayed decimal shows calculation, not known measurement precision. No score is a population percentile, ability verification, diagnosis or job recommendation.

The item, scale and fixed interleaved order are versioned together as 2026-09.1; presentation is continuum-v2, scoring is mean-v2, interpretation is report-v2. The RIASEC report shows exact ties together and describes scores within one response-step as close. That rule is an editorial precaution, not a confidence interval. Ratings across different questionnaires measure different things and must not be directly compared as if they shared a unit.

Old 2026-08.1 results keep their original mean-v1 score and report-v1 interpretation. Their items and five-point response meanings were different, so the revised questions cannot be used to rescore old answers or to claim measured change from the old score. Local drafts are separated by instrument, version, questionnaire language and owner. Saving a result still requires an explicit action and account sign-in.

Language and intended use

Current complete revised item sets are available in English and Chinese; RIASEC also has Spanish, Japanese and Armenian wording, while DISC also has Spanish wording. These translations have not yet completed native-speaker cognitive interviews or measurement-invariance studies. Other questionnaire-language combinations use a complete English form, with a notice in the requested language. The questions, response labels and report use the same form language. Locale is not nationality, and scores from different languages are not established as interchangeable.

These tools support personal reflection and coaching conversations. They are not validated for hiring, admission, promotion or other consequential decisions. Skill confidence is self-perception, not demonstrated ability; users should connect it to work samples and feedback. The color and DISC labels do not imply licensed or validated commercial instruments.

Evidence still needed

The next stage starts with target-user cognitive interviews and independent reviewers with measurement, occupational and language expertise. We need to check whether people interpret each prompt and response option as intended, whether key parts of each construct are missing, and whether socially desirable or opportunity-dependent answers dominate. We then need a consented, versioned pilot with item-level analyses, appropriate reliability estimates, comparison with relevant external measures and language-equivalence checks. Annual retakes are not a substitute for a short-interval reliability study. Completion rate and enjoyment are useful product metrics, not validity evidence.

A context packet uses schemaVersion 1.1 and can carry a nullable dimension score, item count, questionnaire locale and versioned report. It contains no raw item answers. Task and scope grants still control reads, and the policy marks assessment use as exploration-only.