Methodology
What this assessment is, and what it isn't.
The short version
This is a structured framework for comparing career directions against your constraints and what drains you. It is not a validated psychometric instrument. It has no norms, no population to compare you against, and no test-retest data. Nobody has checked whether two people with the same answers get the same result a month apart, because we haven't collected the data to check.
Most assessments won't tell you that. We would rather you knew where the edges are.
What the questions are doing
The assessment asks what drains you rather than what interests you. Self-reported interest is a weak predictor of whether you'll actually be happy in a job; the things that reliably exhaust you are easier to report accurately and harder to talk yourself out of.
It also asks for evidence rather than self-assessment: what you've done for two hours without resistance, what other people praise you for, what you avoid despite being good at it. Forced trade-offs are used where a plain question would invite a flattering answer.
Where the role data comes from
Each of the 52 career directions is mapped to an occupation in the O*NET database (release 30.3), published by the U.S. Department of Labor. The demands each role places on you, meaning how much contact with others, how much time pressure, how much repetition, are read from O*NET's survey data rather than guessed at.
26 of those 52 mappings are approximate. Job titles like DevOps Engineer, Customer Success Manager and Revenue Operations Analyst don't exist in O*NET, and some occupations that do exist carry no survey data. Those roles are matched to the nearest surveyed occupation, which is a real limitation: the further the substitute is from the job you actually want, the less the demand profile tells you.
Which parts are measured and which are opinion
13 of the 19 things we ask about map onto a published O*NET measure. The remaining 6, covering how you feel about deep focus, message triage, first drafts, optimisation, auditing, and maintenance work, have no defensible counterpart in the data, so those parts of the match are authored by us. They are informed guesses, and you should read them as such.
The weights that combine these into a ranking were also chosen by us, not fitted against outcomes. We have no data on whether people who followed the advice ended up better off.
Why fit is a band, not a percentage
An earlier version showed a fit percentage. A number like “73%” implies a precision that authored weights can't support, so results are now grouped into strong, plausible, and stretch. The ordering is the part we stand behind; the band says how much.
A band also accounts for how much you told us. If you skipped most of the assessment, nothing will come back as a strong match, because we don't know enough to claim one. Where two archetypes score close together, both are shown rather than picking a winner and presenting it as a verdict.
Where the salary figures come from
Pay figures are published by the U.S. Bureau of Labor Statistics in its Occupational Employment and Wage Statistics survey. They are US national averages across all industries for the mapped occupation. They are not offers, not predictions, and not adjusted for your location, seniority, or industry. If you are outside the United States they are context at best.
The report is instructed to quote these figures exactly or not at all. Where an occupation has no published figure, the report should say pay varies rather than estimate one.
The report itself
The written report is generated by an AI model working from your answers, your computed results, and the occupational data above. It can still be wrong. It is checked before delivery for length, completeness, and consistency with the preview you saw, but those checks catch broken output, not bad judgement.
Attribution
This site incorporates information from the O*NET Database, version 30.3, by the U.S. Department of Labor, Employment and Training Administration. Used under the Creative Commons Attribution 4.0 International License. O*NET® is a trademark of the U.S. Department of Labor, Employment and Training Administration. Our matching and scoring are our own work and are not endorsed by, and do not reflect the views of, the Department of Labor.
Wage data: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics.
Questions
If something here doesn't add up, tell us at support@idontknowwhatimdoing.com.