causalis.refutation.sensitivity_analysis

causalis.refutation.sensitivity_analysis(effect_estimation, *, cf_y, cf_d, rho=1.0, level=0.95, use_signed_rr=False)

Compute bias-aware components and cache them on effect_estimation[“bias_aware”].

Return type:

Dict[str, Any]

Parameters:

• effect_estimation (Dict[str, Any])

• cf_y (float)

• cf_d (float)

• rho (float)

• level (float)

• use_signed_rr (bool)

Returns a dict with:

• theta, se, level, z

• sampling_ci

• theta_bounds_confounding = (theta - max_bias, theta + max_bias)

• bias_aware_ci = [theta - (max_bias + z*se), theta + (max_bias + z*se)]

• max_bias and components (sigma2, nu2)

• params (cf_y, cf_d, rho, use_signed_rr)