Psychometrician — Scale Development, Validation & Measurement
Turns your agent into a psychometrician that builds, validates, and troubleshoots measures — reliability, validity, factor analysis, IRT, and measurement invariance.
- Apply psychometric standards like McDonald's omega to scale validation
- Diagnose a badly fitting CFA — estimator, global vs. local fit, respecification
- Test configural / metric / scalar invariance before comparing groups or time points
$19
· or 95 creditsSecure checkout via Stripe
Included in download
- Apply psychometric standards like McDonald's omega to scale validation
- Diagnose a badly fitting CFA — estimator, global vs. local fit, respecification
- file_read, file_write automation included
- Ready for Works in Claude Code
Media gallery
See it in action
You say
My 12-item scale has Cronbach's alpha of .74, and my CFA has CFI = .91 and RMSEA = .09. Is the scale fine?
Your agent does
Not yet — two things to fix first. Reliability: alpha assumes equal loadings, which your CFA loadings probably violate, so report McDonald's omega with a confidence interval instead; if the scale is multidimensional, check omega-hierarchical and ECV before trusting a single total score. Fit: CFI = .91 and RMSEA = .09 both sit outside the usual guideline range (CFI around .95+, RMSEA around .06 or lower), so the model likely misfits — read the largest residual correlations and modification indices to see where, and if the items are ordinal, estimate with WLSMV on polychoric correlations rather than ML. Treat any respecification as exploratory and confirm it in a fresh sample.
Psychometrician — Scale Development, Validation & Measurement
Turns your agent into a psychometrician that builds, validates, and troubleshoots measures — reliability, validity, factor analysis, IRT, and measurement invariance.
$19
· or 95 creditsSecure checkout via Stripe
Included in download
- Apply psychometric standards like McDonald's omega to scale validation
- Diagnose a badly fitting CFA — estimator, global vs. local fit, respecification
- file_read, file_write automation included
- Ready for Works in Claude Code
- Instant install
Media gallery
See it in action
You say
My 12-item scale has Cronbach's alpha of .74, and my CFA has CFI = .91 and RMSEA = .09. Is the scale fine?
Your agent does
Not yet — two things to fix first. Reliability: alpha assumes equal loadings, which your CFA loadings probably violate, so report McDonald's omega with a confidence interval instead; if the scale is multidimensional, check omega-hierarchical and ECV before trusting a single total score. Fit: CFI = .91 and RMSEA = .09 both sit outside the usual guideline range (CFI around .95+, RMSEA around .06 or lower), so the model likely misfits — read the largest residual correlations and modification indices to see where, and if the items are ordinal, estimate with WLSMV on polychoric correlations rather than ML. Treat any respecification as exploratory and confirm it in a fresh sample.
About This Skill
A score is an inference about something you can't observe directly, and most measurement mistakes come from treating that inference as if it were the thing itself. This skill gives your agent a psychometrician's judgment for building measures that mean the same thing across people, groups, and time — and for deciding whether an existing measure is good enough for the use you're putting it to. The core frames validity as an argument, with reliability, dimensionality, and invariance as the evidence behind it. Five reference files go deep: reliability in all its forms (omega, ICC, generalizability theory, standard error of measurement), the five sources of validity evidence, factor analysis (EFA, CFA, ESEM, bifactor, network/EGA), item response theory and Rasch, measurement invariance and differential item functioning, and the full scale-development lifecycle with reporting standards. Use it to develop a new scale, validate or adapt an existing one, plan a reliability or invariance analysis, or work out why a factor model is fitting badly. It's calibrated to current practice — omega over alpha, fit indices as guidelines rather than gates, partial invariance and alignment when exact invariance fails — and it keeps cutoffs honest as context-dependent guidance. Includes a source-verified list of canonical references across every subfield it touches. Written for researchers, scale developers, assessment and I/O professionals, clinical and educational measurement teams, and graduate students.
Use Cases
- Apply psychometric standards like McDonald's omega to scale validation
- Diagnose a badly fitting CFA — estimator, global vs. local fit, respecification
- Test configural / metric / scalar invariance before comparing groups or time points
- Lay out a validation plan for a new scale, from item pool to invariance
- Pick the right reliability coefficient (omega variant, ICC model) and report it with a CI
- Screen for careless responding with pre-specified, multi-indicator rules
Known Limitations
Provides methodology and interpretation; it doesn't fit the models on your data for you.
Names and explains established instruments; it doesn't reproduce copyrighted scale items.
Fit and reliability thresholds are guidance to reason from, not automatic pass/fail lines.
How to Install
mkdir -p ~/.claude/skills && curl -sL https://www.agensi.io/api/install/psychometrician-scale-development-validation-measurement -o /tmp/psychometrician-scale-development-validation-measurement.zip && unzip -o /tmp/psychometrician-scale-development-validation-measurement.zip -d ~/.claude/skills && rm /tmp/psychometrician-scale-development-validation-measurement.zipFree skills install directly. Paid skills require purchase - use the download button above after buying.
Reviews
No reviews yet - be the first to share your experience.
Only users who have downloaded or purchased this skill can leave a review.
Early access skill
Be the first to review this skill.
Only users who have downloaded or purchased this skill can leave a review.
Security Scanned
Passed automated security review
Permissions
File Scopes
Pure Markdown guidance — no scripts, network calls, or secrets. It needs read access only to whatever you give it (items, CFA/IRT output, notes); write access is optional for saving a plan or report. No terminal, browser, network, or environment-variable access.
Tags
Works in Claude Code, Cursor, Codex CLI, OpenClaw, and any SKILL.md-compatible agent. Pure Markdown with progressive disclosure (a core file plus reference files loaded on demand); no scripts, network, or setup. Best on frontier models, where the agent reasons over the methodology rather than pattern-matching.
Frequently Asked Questions
Learn More About AI Agent Skills
More Premium Skills
designing-hybrid-context-layers
Architects the right retrieval strategy for every query — teaching your agent when to use RAG, a knowledge graph, or a temporal index instead of defaulting to vector search for everything.
Bounty Security Pattern Master Library — 399 Vulnerability Patterns
A premium library of 399 vulnerability patterns and DeFi attack vectors for AI-driven bug hunting and security audits.
Cinematic Landing Page Builder
Turn any business URL into a high-end animated landing page with 4K AI assets and GSAP animations via Cloudflare.
ai-automation-qa-pack
Professional QA & UAT documentation generator for AI automation agencies and complex agent deployments.