OutcomeFalsification.run#

OutcomeFalsification.run(experiment, context)[source]#

Run outcome falsification analysis.

For each falsification formula, fits the experiment with the alternative formula and reports the estimated effect size with HDI intervals.

This is an informational check (passed=None). The researcher interprets whether the pattern of effect sizes across outcomes supports their causal story.

Parameters:
  • experiment (BaseExperiment) – The fitted experiment to check.

  • context (PipelineContext) – Pipeline context providing experiment_config for re-fits.

Returns:

With passed=None (informational). The table contains effect sizes and HDI intervals for each falsification formula.

Return type:

CheckResult