ConductCognition test census: use and evidence data
Release 1.0.1, 8 November 2026
Page: https://conductcognition.com/data

Source: ConductCognition test census 0.22.0, counted 31 October 2026, 2 November 2026 and 3 November 2026 from OpenAlex and ClinicalTrials.gov.
Registered studies were read from ClinicalTrials.gov on 8 November 2026, limited to studies first posted by each family's count date.

ROWS
Each row is one test family, named by the ConductCognition tests counted in it: 451 tests in 447 families. The census counts 495 families. The ones outside the library are not named; concentration.json lists them as not-in-library-<rank> with their domain and papers per year, so every figure can be rebuilt. Counts describe the published research on each task, not ConductCognition's version of it.

FILES
use-and-evidence.csv: One line per test family on this page: every column of the table, the registry splits, catalogs, flags and papers per year.
details.json: Each family's registered studies by outcome list, condition group, start year, phase, type and sponsor, its search queries and its catalog entries.
registry-studies.csv: One line per registered study and family: NCT ID, outcome list, condition groups, phase, study type, sponsor class, start year and enrollment.
queries.json: The exact OpenAlex and ClinicalTrials.gov searches behind each family's counts.
concentration.json: The concentration figures and their inputs: papers per year for every census family, those outside the library without names.
README.txt: What each file holds, the release history and how to cite the data.

DEFINITIONS
Papers: OpenAlex works whose title or abstract names the task, works whose main topic is in the physical sciences left out.
Test-retest papers: papers from any year that also mention test-retest. Reliability or validity papers: papers that also mention test-retest, reliability, validity, validation, psychometrics or normative data.
Registered clinical studies: ClinicalTrials.gov studies that name the task in an outcome measure. Primary endpoint studies name it in a primary outcome. A study whose outcome text the plain check cannot place is counted as unclear.
Condition groups: matched on the study's condition names and MeSH terms. A study can fall in several groups; one in none is 'other'. The top condition is the named group with the most studies.
Census addition: a family that none of the six task catalogs the census reads lists (Cognitive Atlas, Inquisit library, PsyToolkit, PEBL, Experiment Factory, Gorilla sample tasks).
Search precision: the share of matched papers that really use the task. Not yet audited.

HOW TO CITE
ConductScience. (2026). ConductCognition test census: use and evidence data (Version 1.0.1) [Data set]. https://conductcognition.com/data

@misc{conductcognition_census_2026,
  author       = {{ConductScience}},
  title        = {ConductCognition test census: use and evidence data},
  year         = {2026},
  version      = {1.0.1},
  howpublished = {\url{https://conductcognition.com/data}},
  note         = {Data set, test census 0.22.0; registered studies read 8 November 2026}
}

RELEASE HISTORY
1.0.1 (8 November 2026): The page and its files moved to https://conductcognition.com/data, the files to /data/download. No figure changed.
1.0.0 (8 November 2026): First public release: the census counts, registered studies by outcome list, condition, start year and phase, concentration figures, search queries and catalog crosswalk.