Methodology — How the Tests Are Written and Scored
This page collects the rules every test, quiz and tool on aimbst is written by. They are the same rules we have followed since the site started; putting them in one place lets you check our claims against any page you are on.
We only build on public models
When a questionnaire test needs a theoretical backbone, we use models that are in the public domain or explicitly free for commercial use. Career interests follow Holland's RIASEC framework. Personality traits follow the IPIP item pool tradition for the Big Five. Knowledge quizzes follow public scoring conventions. The model is named on the page that uses it.
The questions themselves are always written from scratch for this site. Using a public model means borrowing the map, not the furniture: the framework tells us what to ask about, and we write how to ask it.
Sources International Personality Item Pool (IPIP) O*NET Interest Profiler — Holland's RIASEC IPIP — “Multi-Construct IPIP Inventories”, listing the Lexical Big Five Inventories and Goldberg’s (1992) Big-Five Factor Markers
What we refuse to do
We do not reproduce items from published or commercial assessments — not reworded, not translated, not paraphrased. We do not use trademarked instrument names to describe our tests. And we do not use scales whose licenses forbid commercial use: that is why you will not find a burnout inventory, a sleep-apnea screener or the classic color-plate test here, even though they would be popular.
The medical category as a whole is on hold. A website quiz cannot examine you, and a wrong reassurance is worse than no answer. Tests that touch psychology carry a plain disclaimer saying they are not a diagnosis — and they mean it.
How scoring works
Questionnaire tests score in your browser, from your answers only. Each answer moves one or more named scales; some questions are reverse-scored so that agreeing with everything does not produce an extreme profile. The result you see is the scale pattern your answers drew — which is why answering the same way twice gives the same result. There is no randomness and no hidden weighting.
Knowledge quizzes have fixed answer keys, printed on the page below the quiz so you can check every answer against the questions. Hardware tools have no answer key: they read what your device or your own timing reports to the browser — key events, mouse events, audio levels, screen properties — and show it back to you beside the usual range.
What never leaves your browser
Your answers, your microphone and camera streams, your key presses and your results are processed in the page and are never uploaded to or stored on a server. The microphone test meters your input level, and the short sample it can record for playback stays in the page; the webcam test draws your camera to your own screen and nowhere else; the typing test counts your keystrokes in the page. Close the tab and it is gone.
The two small exceptions live in your own browser, not on our servers: a daily-streak counter and your zodiac sign choice are kept in your browser's local storage so the page can greet you correctly tomorrow. The privacy page lists both.
Before a page goes live
Nothing here publishes itself. Every test and guide moves through the same fixed sequence: choosing the subject, writing the questions, localising them, building the page, then checking the result. A person decides both what gets built and when the site is published, and the publishing step needs an explicit instruction each time.
A set of automated checks runs on every build, and the build stops if any one of them fails. They verify the things this page promises. Tests touching psychology must carry their disclaimer. No page may contain a way to send your answers anywhere. Result pages must stay out of search, so the test page remains the entry point. No images may creep in, and pages must stay within the weight the front page claims. Separate unit tests cover the scoring logic itself, so a result cannot silently change.
The checks cannot judge everything. What gets built, and whether a page is ready to be seen, is decided by a person — publishing takes an explicit instruction every time, so nothing reaches you without that decision. The craft in between, such as whether a translation reads the way a native speaker would write it, is worked out in-house as the page is written, held to the rules on this page. We would rather tell you where the line sits than claim a pair of eyes on every sentence.
No images, by design
There is not a single photograph or image file on these pages. Every visual — color plates, world maps, pianos, dials — is drawn with code (SVG and CSS). This is why pages stay under a fraction of the weight of a typical web page and load quickly on slow connections. It also means every visual scales cleanly on any screen.
The exceptions sit outside the page itself, and they prove the rule: the small icon your browser tab shows, and the preview card messaging apps want when someone shares a link, for which this domain does hold a set of PNG files. Those cards never appear on any page — nothing here loads them — and a Python script draws them the same way the pages are drawn, filling a rectangle and setting type. No stock photography, no image generator, no camera.
If you find a mistake
Quiz answers are checked against current facts when written and revisited when the world changes — but mistakes happen. If an answer looks wrong or a translation reads badly, the contact page has a working email address. Corrections ship in days, and the page's revision date changes so you can see it happened.