Data & analysisExperiment qualitydata

Check an experiment's assignment before reading outcomes

Tests randomisation balance, exposure and exclusions independently of results.

Editorial illustration for this prompt; the prompt itself does not generate this image.

Ready to use

Prompt

Using [ASSIGNMENT LOG], [EXPOSURE LOG] and [EXPERIMENT PLAN], audit whether participants were allocated as planned before calculating outcome differences. Report counts by group, eligibility, duplicate assignments, crossover, missing exposure and important baseline imbalances. Provide formulas, a diagnostic table and a stop/go recommendation limited to data integrity. Do not peek at outcomes to redefine groups or invent statistical significance. Check timezone boundaries and whether exclusions were pre-specified. Mark privacy-sensitive attributes for approved aggregation rather than displaying individual records.