A Mental Model

A certification is a living thing.

One standard sits at the center. Four rings of authority surround it. Five capabilities orbit as spokes. Decisions travel outward, evidence travels home, and the whole map turns, because iteration is the climate.

The Distinction
The building code, and the blueprint
The Body of Knowledge governs the standard. The Learning Architecture operationalizes it. Hover either card to open what each one is.

Body of Knowledge

The Building Code

The written framework that defines what a credential proves and how it stands up to scrutiny.

Hover to open ▾
  • Defines what the certification guarantees and the standard it defends.
  • Governs scope, evidence rules, and the logic of a passing score.
  • Serves as the single source of truth for leadership and auditors.
  • Aligns to external standards such as ISO/IEC 17024.
  • Holds steady while the content beneath it evolves.
  • Answers one question: what does this credential mean.

Learning Architecture

The Blueprint

The visual and working structure that turns the standard into a path a learner can travel.

Hover to open ▾
  • Operationalizes the standard into teachable, sequenced structure.
  • Maps work domains, roles, duties, and evidence by level.
  • Charts the route from readiness to demonstrated mastery.
  • Lives as the visual map and the working workbook.
  • Guides designers, SMEs, and stakeholders through the build.
  • Answers one question: how this credential is built and navigated.
Because the Body of Knowledge governs, the two agree by design.
The Capability Ladder
The three certification levels
Each level rises in cognitive demand, from running the work, to deciding it, to designing it.

Technician · L1

Role: Face of the business, Tier 1.
Peak: Apply.
Verbs: choose, solve, prepare.

Administrator · L2

Role: Hinge of the ladder, certified core.
Peak: Evaluate.
Verbs: appraise, judge, defend.

Expert · L3

Role: Architect of the practice.
Peak: Create.
Verbs: design, formulate, originate.
Methodologies
Models I command, and when I choose them
A methodology fits the case. These are the models I reach for, each with what it is and when it earns the call.

ADDIE

Structured spine

Analyze, Design, Develop, Implement, Evaluate. The disciplined backbone that keeps every stage traceable from intake to launch. I choose it when a program needs rigor, auditability, and a stable structure.

SAM

Agile prototyping

The Successive Approximation Model builds through rapid prototypes and continuous feedback across successive cycles, so the design sharpens with each pass. I choose it when speed matters and the shape emerges as we build.

Kirkpatrick

Four levels of impact

The evaluation model that measures whether learning reached real work. I choose it to prove a program changed behavior and delivered results.

  • 1Reaction: how learners respond, their engagement and sense of relevance.
  • 2Learning: the knowledge, skill, and confidence gained.
  • 3Behavior: how learners apply the learning in real work.
  • 4Results: the business outcomes that follow.
How it proves · a measurement science
Test the reasoning beneath the click
What psychometrics is

Psychometrics is the science of measuring the invisible. It turns abstract capability, what a person knows and can do, into objective evidence through standardized instruments and statistical models. A well built exam becomes a measuring device, and psychometrics proves the device measures truly and fairly.

Standard Setting

The Angoff method. Eighty percent as the provisional bar, refined by SME panels into a defended cut score.

Each expert judges the probability that a minimally qualified candidate answers each item. The averaged judgments set a passing score the certification can defend to an auditor.

Reliability

Internal consistency at 0.80 or higher, proving the exam measures capability across forms.

α =kk − 1(1 −ΣViVt)
Cronbach's alpha reads the share of a score that reflects true capability against the share that reflects noise. A value at 0.80 and above signals a form that holds together. Cronbach's alpha reads scales, and KR-20 reads correct or incorrect items. The demonstration computes KR-20.

Item Health

Difficulty lands between 30 and 90 percent. Point-biserial reaches 0.20 or higher to discriminate.

Difficulty confirms an item sorts the cohort. Point-biserial confirms the strong candidates answer it correctly more often, so each item earns its place on the form.
The standards the exam answers to: ISO/IEC 17024 for fairness and defensible scores, NIST and NICE for the vocabulary of tasks, knowledge, and skills.
Proof in motion
One hundred learners meet the standard
Illustrative model · synthetic data
A generated cohort of one hundred, spread across novice, proficient, and expert ability. The exam carries weighted domains, the cut score sets the bar, and the reliability reads live. The method is real; the learners are generated. Move the cut score, and watch standard-setting decide who passes.