Career direction test Not a list of occupations but a decision you make yourself

The core of the test is six interest types on which career guidance has run for half a century: realistic, investigative, artistic, social, enterprising and conventional. Your three leading types form a profile code, and it is that code the algorithm matches directory candidates against. This is a fundamentally different object from a four-letter type: a type is a label that sorts people into sixteen boxes, while the hexagon is the shape of a profile, and it has no correct pole. A low realistic type is not a shortcoming: it is the knowledge that a skilled trade with a short entry will cost you more than it costs others — knowledge better obtained before a retraining course than after one.

For an adult this foothold is firmer than for a schoolchild, and that is the only genuinely strong argument against the common «career guidance is for children». Interests stabilise sharply at around twenty-two and then hold for two decades more reliably than personality traits do. Which means an adult in doubt stands on firmer ground than a final-year pupil taken to a career session with the whole class.

• Interests — six types, giving the hexagon, the profile code and the shortlist of directions;
• Work values — achievement, independence, recognition, relationships, support and fairness, conditions and stability: what the work is for;
• Work-related traits — organization, stress sensitivity, working with people, appetite for the new;
• Your current work — tasks, people, meaning and the job as a whole: what exactly is wrong in it;
• What blocks the choice — six different difficulties, each with its own answer;
• Readiness to move — concern for the future, sense of control, curiosity and confidence that you can manage;
• Your constraints — runway, room for income to drop, room to study, mobility, backing from those close;
• How you are doing — exhaustion, mood background and well-being: the branch that overrides every other.

Information about the world of work is one of the five components whose presence decides, in the research, whether a career intervention works at all. The trouble is that a ready mapping from interest code to occupation exists for the US labour market, and half of those positions do not exist here while half of the local ones are missing there. So the directory is our own: 192 occupations with an interest code, leading work values, the education expected, whether a common route in without a matching diploma exists, how remote the work can be, whether the direction typically appears in free state retraining lists, and an assessment of how age-open the entry is.

The matching is done by algorithm: it scores how well your profile fits each occupation’s code, applies your constraints, and hands the reading a short list — no more than three positions from one sector, so that the whole list does not collapse into a single industry. The model receives that list ready-made and only explains why it looks the way it does. It cannot invent an occupation, a salary or a vacancy count: none of that is in what it was given. Pay is deliberately absent from the directory — it changes faster than any report, and naming it would be lying to within a month.

The commonest reason a well-weighed decision stays unmade is not a lack of motivation but a lack of money for a pause. Tests do not ask about it because a list of occupations is indifferent to your circumstances. A reading is not. There are five scales here: whether there is a cushion for at least a few months, whether the income may fall or other people depend on it, whether there is energy and time to study after the working day, whether the schedule, format or city can change, and whether there is anyone to lean on.

It then works literally: if there is no cushion and the income cannot break, the algorithm lowers everything that requires a long entry, and the reading builds only routes that start alongside the current job. No recommendation here begins with the word «quit». And where a constraint and a wish genuinely conflict, the reading names the conflict and its price out loud rather than passing over it politely: a horizon with a real first step beats a fantasy with a deadline.

The conclusion here forks rather than running on one track, and the fork is computed by algorithm. «State first» — when there is nothing left: in that state nothing looks interesting, and matching an occupation answers the wrong question. «Reconnaissance first» — when what is missing is information rather than resolve: the answer is then three concrete steps (a conversation with someone inside the occupation, a trial task, a day of observation), not «think it over some more». «A change is justified» — when there is a pull away and a sense of where to; the first alone is not enough, that would be flight. And «reshape the current job» — when the job broadly holds and one specific part has sagged.

That last answer deserves a separate word. Reshaping tasks, relationships and meaning inside the same role is linked in research to higher satisfaction and lower burnout without any change of occupation. It is an honest negative conclusion, and structurally no free test that lives off selling courses can give it: telling someone «change nothing» means not selling them any training.

The link between interest fit and how content people are at work exists, but it is moderate: across more than a hundred studies it explains about one part in twenty-five of the differences between people. With how well someone performs the link is noticeably stronger. But a manager, a workload, money and health still weigh more than a job title.

We say this here, on the page, rather than in small print at the end of a report — because what it is honest to promise depends on that figure. Selling you «your occupation» on that basis would be a lie. So what comes out here is directions, the price of a move and first steps, and the decision stays yours. And it is the ability to decide, not the accuracy of a match, that shifts the most in career-intervention research.