Skip to content

Learning Design Wiki

A persistent, LLM-maintained knowledge base for learning design. Read CLAUDE.md for the schema, page templates, and agent operating instructions.

Read this first

  • Models wrote this wiki, and no person has yet checked any page. Every page's Page metadata panel shows its trust tier; today all of them read unverified. Treat a page as a well-sourced lead, and check the cited study before relying on a detail.
  • Most claims rest on one study. A claim page reports what one or a few studies found, in the conditions they ran. Where several studies bear on one question, a table on the claim page lists them one row each; nothing is averaged into a verdict.
  • Every study carries codes. q1–q4 is the design tier (q4: a pre-registered trial or well-powered meta-analysis; q1: a case study or argument). i1–i3 is the size of a printed effect (small, medium, large); i? means the source printed none, which is not the same as no effect. causal · r2 names the kind of evidence and how well it was done for its kind (r1–r3). The evidence report shows the state of it all.
  • Markers on design pages say how a claim bears on the page: + supports, ~ depends on conditions, - counts against, and S/M/W how strong, capped by the claim's recorded evidence.
  • Principle and pattern pages say what is tested and what is not. A design marked Default design, while the relationship is untested is a labelled proposal, not a finding.
  • Found an error? Open an issue on GitHub with the page and the source.

Knowledge Types

Principles (170)

Research-backed design commitments: what to do and why.

Elements (1057)

Instructional building blocks — the components you compose into patterns.

Patterns (54)

Reusable instructional designs at the lesson or unit level.

Designs (1022)

Designs for a particular setting, course, population or product: what a design uses the general patterns for. A pattern is general across contexts; a design is not.

Products & Programmes (1919)

Named products and programmes others adopt: software and platforms, curricula, assessments and instruments, datasets, and branded programmes and initiatives, each with its own frameworks and the evidence about it. A design is one setting's.

Strategies (6733)

Concrete teaching activity recipes — specific, implementable approaches.

Design Processes (18)

How a course gets designed — whole-process models a designer works through, rather than anything a learner meets.

Design Methods (37)

The practices a design process is made of — analysis, elicitation, mapping and evaluation work done on the design, not in the classroom.

Research Methods (641)

Methods for studying learning and evaluating education that are specific to it or especially useful in it. General social-science methods are not listed.

Theories (1360)

Explanatory frameworks that ground principles and claims.

Learner Variables (75)

Canonical learner characteristics claims report findings about — one page per variable, so the same concept does not fragment across differently-worded tags.

Claims (26894)

Empirical claims with evidence ratings, sources, and competing views.

Quick navigation

How to use this wiki

As an agent: read CLAUDE.md first. Use index.md as your entry point, follow the markdown links in each page to traverse the graph, and use grep for targeted search.

As a human: browse the folder indexes above, or search from the box at the top of every page (on GitHub, read the wiki at wiki.learningdesignalliance.org). Evidence tags ([+S], [+M], [~M], [-W]) indicate claim support strength. Pages marked status: draft are stubs or unreviewed; review pages have content that needs an expert's check.