causalis.data_contracts.multi_gate_estimate

Joint subgroup effect results for multiple treatments versus one baseline.

Module Contents

Classes

MultiGateEstimate

Group-by-treatment effects with pointwise inference and joint covariance.

API

class causalis.data_contracts.multi_gate_estimate.MultiGateEstimate(/, **data: Any)

Bases: pydantic.BaseModel

Group-by-treatment effects with pointwise inference and joint covariance.

Effect arrays have shape (len(group_names), len(treatment_names)). treatment_names contains only active arms. Covariance axes use the group-major (group, treatment) product. Joint covariance does not imply simultaneous confidence intervals or multiplicity-adjusted significance.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

model_config

‘ConfigDict(…)’

estimand: str

None

model: str

‘MultiTreatmentIRM’

group_names: List[str]

None

treatment_names: List[str]

None

baseline_treatment: str

None

values: numpy.ndarray

None

std_errors: numpy.ndarray

None

test_stats: numpy.ndarray

None

p_values: numpy.ndarray

None

ci_lower: numpy.ndarray

None

ci_upper: numpy.ndarray

None

alpha: float

None

covariance: pandas.DataFrame

None

summary_table: pandas.DataFrame

None

n_group: numpy.ndarray

None

n_treated: numpy.ndarray

None

n_control: numpy.ndarray

None

n_other: numpy.ndarray

None

model_options: Dict[str, Any]

‘Field(…)’

diagnostic_data: Optional[Dict[str, Any]]

None

time: str

‘Field(…)’

summary() → pandas.DataFrame

Return one row per group and active treatment, with pointwise inference.

for_treatment(name: str) → causalis.data_contracts.gate_estimate.GateEstimate

Extract an arm for existing subgroup contrasts and gate_plot.

For GATET, each arm targets its own recipients. Differences between arms’ GATETs therefore also reflect different target populations.