00

A test everyone passes measures nothing.

The standard test at the end of a module has four questions whose correct answer jumps straight out at you. Everybody passes, the record turns green, and the organisation believes the matter is settled. We build assessment that actually tells you something, and records you can put in front of an auditor.

Ask for an example from your sector

Technical specs

Levels
Knowing, deciding, doing
Records
Per person and per version
Integration
SCORM 1.2 / xAPI
Lead time
From 3 weeks
01 How it works

We separate three levels. Knowing is tested with knowledge questions, which is necessary but rarely sufficient. Deciding is tested with situations where several options are defensible and consequences follow. Doing is tested at the workplace, through observation, and that cannot be solved digitally.

The quality sits entirely in the wrong answers. A question with one plausible and three nonsensical options tests whether somebody can read. Distractors an experienced colleague could also defend come out of incident reports and out of conversations with your own specialists.

Often this starts not with building but with clearing out: going through the existing tests, removing the questions that measure nothing, and establishing which operations genuinely require observation. That is less work than starting over and usually delivers more.

What we set up

  • Scenario tests with credible distractors
  • Observation forms for assessment at the workplace
  • A record per person, including which version applied
  • Monitoring of validity periods and refresher instruction
  • Reporting that shows which mistake recurs across a whole shift
02 When to use

Needed

Work where authorisation has to be demonstrable. Operations where a mistake is expensive or dangerous. Organisations with certification periods or external audits.

Not needed

Information everybody only has to be aware of, with no consequences attached. Subjects where a conversation with the supervisor delivers more than a score.

The most underused return on assessment is not the verdict on the participant but the measurement of your own material. When eighty percent get the same question wrong, that is not a knowledge problem in eighty percent of your people.

04 Specifications

Audit-proof

Recorded per person: what was completed, when, and which version applied.

Starts with what exists

Clear out the existing tests first, only then build what is missing.

At the workplace too

Observation is part of it, because not everything can be measured behind a screen.

Do you need to demonstrate who is competent for what? We can help with that.

M

Milan

Business development, STARK Learning