InversePropensityWeighting.effect_summary#

InversePropensityWeighting.effect_summary(*, window='post', direction='increase', alpha=0.05, cumulative=True, relative=True, min_effect=None, treated_unit=None, period=None, prefix='Post-period', **kwargs)[source]#

Generate a decision-ready summary of causal effects.

Note

This method is not yet implemented for InversePropensityWeighting experiments. Calling it will raise NotImplementedError.

Parameters:

  • window (Union[Literal['post'], tuple, slice]) – Time window for analysis. Defaults to "post".

  • direction (Literal['increase', 'decrease', 'two-sided']) – Direction for tail probability calculation. Defaults to "increase".

  • alpha (float) – Significance level for HDI/CI intervals. Defaults to 0.05.

  • cumulative (bool) – Whether to include cumulative effect statistics. Defaults to True.

  • relative (bool) – Whether to include relative effect statistics. Defaults to True.

  • min_effect (float | None) – ROPE threshold for practical equivalence. Defaults to None.

  • treated_unit (str | None) – For multi-unit experiments, the unit to analyse. Defaults to None.

  • period (Optional[Literal['intervention', 'post', 'comparison']]) – Period to summarise for multi-period experiments. Defaults to None.

  • prefix (str) – Label prefix for prose generation. Defaults to "Post-period".

  • **kwargs (Any) – Additional keyword arguments (currently unused).

Raises:

NotImplementedError – Always raised; this method is a placeholder for future work.

Return type:

EffectSummary