Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Switch base URL to api.adaptivetest.io custom domain
What This Skill Does
Covers the design and implementation of computerized adaptive tests (CAT) using Item Response Theory. Includes IRT models (1PL, 2PL, 3PL), item selection algorithms, ability estimation methods, and stopping rules. Applies to K-12 placement, certification, diagnostic, and progress monitoring assessments.
Adaptive tests achieve comparable precision to fixed-length tests with roughly half the items, shortening test time without sacrificing measurement accuracy.
When to use it
- Building a math placement test that adjusts question difficulty in real time
- Calibrating an item bank for a certification or licensure exam
- Estimating student reading level with fewer questions than a fixed-length test
- Setting content balance constraints across multiple subject areas
- Choosing between fixed-length and standard error stopping rules for a diagnostic assessment
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: design a K-12 math placement CAT with 300 items, SE < 0.35 target, max 25 questions, and content balance across algebra, geometry, and statistics
- 1Define test specifications: purpose (placement), ability range (-2 to +2), max length 25 items, SE target 0.35
- 2Specify item bank distribution: 300 items spread across difficulty range with 40% algebra, 30% geometry, 30% statistics
- 3Select algorithms: Randomesque (top 5) for item selection, EAP for ability estimation, combined stopping rule
- 4Outline simulation plan: run 1000 sessions with known abilities to check average test length and item exposure rates
- 5Define output format: ability estimate plus placement recommendation
A complete CAT design spec with item bank requirements, algorithm choices, and stopping rules that places 95% of students within 0.5 grade levels of true ability in an average of 18 questions