Skip to main content

What is the CAT test

I open every certification with a Computerized Adaptive Testing evaluation — CAT for short. It is the only way I can build you a roadmap that matches what you actually know, instead of what you think you know.

How CAT differs from a regular quiz​

A regular quiz hands you a fixed list of questions in a fixed order. CAT does not. After each answer, ARIA recalculates her confidence in your knowledge of every domain on the certification, then chooses the next question to close the biggest gap in that confidence.

The result: I converge on your real skill level much faster than a static quiz would, and I do it without wasting your time on material you already know.

Difficulty adapts per answer​

Every question I ask is tagged with a difficulty from 1 to 5:

  • 1 — Foundational
  • 2 — Applied
  • 3 — Analytical
  • 4 — Complex
  • 5 — Expert

I start you near the middle. From there:

  • Answer correctly, and the next question moves up one level.
  • Answer incorrectly, and it drops one level.
  • The level is clamped to 1–5, so I will never ask you something below the floor or above the ceiling of the certification.

The badge at the top of the question card shows the current level. Read it as feedback — if you are seeing "Complex" or "Expert" questions, I am no longer in doubt about your strong domains.

How long the test runs​

I do not run a fixed number of questions. The evaluation stops when one of these is true:

  • I have asked at least 15 questions AND my overall confidence in your scores has reached 95%.
  • I have asked 25 questions — the hard ceiling, regardless of confidence.

In practice, evaluations finish in 15–22 questions for most users.

The timer is soft​

Each question has a soft per-question pacing target around 60 seconds. I do not auto-submit if you go over — but slow down too much across the whole test and you risk noisy data on the calibration slider, which weakens the signal I use to build your roadmap.

If you genuinely do not know an answer, pick your best guess. CAT is built for that. A confident wrong answer tells me something useful. A drawn-out maybe tells me almost nothing.

You cannot go back​

CAT is sequential — every question I ask depends on every answer before it. If I let you change a previous answer, the entire chain of follow-ups would no longer be valid. So once you submit, the answer is locked.

warning

I will not show you whether each answer was correct during the evaluation. This is an assessment, not a study session. Feedback shifts your strategy mid-test, and that biases the result. You see the full breakdown on the results screen.

What I am building underneath​

While you answer, I track per-domain coverage and per-domain confidence on the server. Both update silently after every question. When I stop the test, I freeze those numbers into your domain scores and levels — and that frozen snapshot is what I use to generate your roadmap on the next screen.