Browse 269 paradigms
Grouped by cognitive domain, each with its description, key references, and every indexed implementation.
23 domains
A map of measurement claims
Every task here is catalogued. Almost none of it has been checked.
This is a catalog of experimental psychology paradigms and the implementations that claim to run them — and a running record of how much scrutiny each claim has actually had. Two kinds of scrutiny, kept apart on purpose.
Human review
0 / 11
specifications reviewed by a named expert.
Every specification in the catalog is sourced: drafted, referenced,
and unreviewed. Raising that state takes a person; no machine check can do it.
the wall
Machine checks
7 / 9
audited implementations carry the contrast they claim.
Each was driven end-to-end by a responder with a known policy. The 2 that failed look like the paradigm and are not wired as it.
The wall is the point. A machine check confirms a task is wired the way a spec says.
It cannot tell you the spec is the right way to measure the thing. That judgement is human, so it is
tracked separately and never inferred from a passing test.
Open call
The first expert review is the one that matters
The catalog's value is not that it lists things — software can list things. It is that a named person, on the record, has judged whether a measurement claim holds. Nobody has done that here yet, and the site says so on every page rather than dressing a green check up as review.
The ask is bounded: one specification, three separable claims, an afternoon. Sign-offs are named, scoped, dated, and revocable.
Start here
Search the corpus
One box across paradigms, specifications, and 994 implementation records.
Runs in your browser
Read the specifications
What it means to implement a paradigm correctly: the defining contrast, the design parameters that move results, and an executable test.
11 specs · 9 audited instances
Compare platforms
Where to build and host: open-source and commercial platforms, timing behaviour, pricing, and the large task collections.
21 platforms
What this catalog does not claim
A link that loads is not a task that works
994 records survive the headless launch check; 148 failed it and are held back. Neither result says anything about the science.
Paradigm labels carry a classifier's error rate
388 records are not confidently mapped to any paradigm and stay out of the paradigm pages. Of those that are mapped, the one confirmed error found so far was a mislabel, not a broken task.
Specifications are drafts
All 11 are unreviewed. They are useful as drafts and dangerous as authority, which is why their state is printed next to every one of them.