A career quiz that explains its answers

The honest job of a career quiz is generating candidates. Most people can name maybe thirty jobs. The U.S. Department of Labor’s occupational database describes hundreds more, and a large share of them are invisible to anyone who has not already met somebody doing them. A quiz that puts eight occupations in front of you that you had never considered has done something genuinely useful, even if you reject seven of them.

The dishonest job is issuing a verdict. No twenty-question instrument knows what you will care about in six years, what your family situation will allow, or which of your interests will survive contact with the actual work. Any tool that speaks as though it does is either overselling or has not thought about it.

So the right posture toward a result is: this is a list to investigate, ranked by something. The only question that matters next is ranked by what, exactly.

Four failure modes account for nearly all of it, and you can spot every one of them from the results screen.

Five things. They are not exotic, and a tool that refuses any of them is one you should be reading skeptically.

The comparison that survives scrutiny is a shape comparison. Describe yourself and describe the occupation on the same set of axes — the same skills, the same kinds of knowledge, the same work activities — and then ask how similar the two profiles are as patterns, not as totals.

There is one correction that makes or breaks this, and it is worth understanding because it explains the "everything scores high" problem. Any two occupations, picked at random, look fairly similar: nearly all work involves some reading, some communicating, some judgment. Raw similarity between two arbitrary jobs sits somewhere around 0.70 to 0.85 before you have learned anything. If a tool reports that number, everything looks like a match and the ranking is noise.

Subtracting the average profile first — centering — throws away the part that every job shares and compares only what makes this occupation distinctive against what makes you distinctive. That is the number worth ranking on, and it is why a well-built quiz produces results that actually spread out.

"Would I like this work" and "could I get this job right now" are different questions with different evidence behind them. The first is about disposition. The second is about what your coursework, projects and résumé can currently prove.

Averaging them produces a single number that answers neither question, and it is the reason so many results feel simultaneously flattering and useless. A seventeen-year-old is not ready for anything yet; that is not a statement about fit. Keep them apart and both numbers become actionable — one tells you where to look, the other tells you what to go build.

Ten questions, about ninety seconds. Your answers build a profile across 161 dimensions taken from O*NET, the U.S. Department of Labor’s occupational database, and that profile is compared against 780+ real occupations described on those same 161 axes. Same axes on both sides, so the comparison is like-for-like.

The score is the mean-centered correlation described above. Fit only — readiness is scored separately and never folded in. Every match opens to show which dimensions pushed it up and which pulled it down, so if you disagree with a result you can see exactly where it came from and decide the tool is wrong. That is the intended use.

The top three matches are visible before you make an account. Each one links to a full page on that occupation — Software Developers, Registered Nurses, Accountants and Auditors and the rest — covering what the work involves, what it takes, and how people get in.