cleverly.validation.ReplicationRecord#
- class cleverly.validation.ReplicationRecord(replicate: int, seed: int, estimand: str, truth: float, estimate: float, std_error: float, covered: bool, rejected: bool, inference_estimate: float, alpha: float, inference: Literal['influence_curve', 'working_mechanism_plugin', 'generated_design_plugin', 'estimated_weight_plugin', 'adaptive_baseline_plugin', 'no_cluster_variance_plugin', 'few_cluster_plugin', 'constant_node_plugin', 'stack_failure_plugin'] = 'influence_curve', support: str = 'experimental')[source]#
One estimand from one successful repeated-sampling draw.
- Parameters:
- replicate
int Index of the draw.
- seed
int Seed that produced it.
- estimand
str Alias this record is for.
- truth
float The estimand’s true value under the process.
- estimate
float What the estimator reported.
- std_error
float The standard error it reported.
- coveredbool
Whether the interval contained
truth.- rejectedbool
Whether the null was rejected at
alpha.- inference_estimate
float The estimate on the inference scale.
- alpha
float Significance level the interval was built at.
- inference
str Which quantity
std_error,coveredandrejectedmeasured. One ofInferenceStatus, which the inference reference lists."influence_curve"says the estimator’s own reported inference. Any other status says the estimator supplies none, and the study measured its retained plug-in diagnostic instead.- support{“supported”, “experimental”}
The support label of the replicate’s estimate.
- replicate