How we assess

We don't examine.
We look at the work.

Multiple-choice tests measure recall, under conditions that never occur again in a professional's life. Bloom assessment is built on a different premise: the only honest evidence of competence is real work, visibly improved, reviewed by a human who knows what good looks like.

Evidence over examination

What conventional testing produces

  • A score that certifies one sitting, on one day, under exam conditions
  • Questions with answer keys — which AI can now produce instantly, making recall a weak signal
  • An incentive to perform the test rather than learn the discipline

What Bloom assessment produces

  • Artifacts — prompts, analyses, revisions, working systems: things you made that exist after the course ends
  • Process visibility — original, diagnosis, revision, result. The improvement is the evidence
  • Human-in-the-loop review — rubric-based feedback from instructors, with revision invited rather than penalized

The improvement is the evidence

1

Original

Your first attempt, preserved — not hidden. Everyone starts somewhere.

2

Diagnosis

What's wrong with it, named precisely — by you, with the frameworks you've learned.

3

Revision

The fix, applied deliberately. This is where technique becomes judgment.

4

Result

The improved work, with the path that produced it fully visible.

A right answer proves little. A visible improvement proves understanding.

The same standard at every tier

Sprints

Each seven-day sprint ends in an artifact — a piece of verifiable work, not a quiz score. The micro-credential certifies the artifact exists and met the standard.

Certificates

Coursework builds a portfolio: assessed assignments with rubric feedback, revision cycles, and a growing body of work that demonstrates the discipline in use.

Capstones

Mastery programs culminate in extended capstone work — projects substantial enough to discuss in an interview, defended rather than merely submitted.

This is harder. That's the point.

Artifact-based assessment is more expensive to run than auto-graded tests. It requires human reviewers, rubrics that hold up under disagreement, and the patience to read real work instead of scanning answer sheets.

We do it anyway, for one reason: a credential is a promise made on your behalf to someone who hasn't met you. We are not willing to make that promise on the strength of a multiple-choice score — and in an era when AI can answer any quiz, neither should you be.

Assessment shows what you can do.

The Reliability Standard shows what you reliably do — dependability observed over time, after the course ends. The same philosophy, extended past graduation: observed, not claimed.

The Reliability Standard