Function for standard deviation and p-value for SAT

[Shicheng-Guo]

Dear Team,

I am curious do we have Function to calculate standard deviation and p-value for ate, something like below:

Predict treatment effect using fitted estimator

ate = meta_reg.predict_ate(X) 

Compute standard errors, confidence intervals, and p-values

cov = meta_reg.covariance_
se = np.sqrt(np.diag(cov))
ci = meta_reg.confidence_intervals_
p = meta_reg.p_values_

Best,

Shicheng

[jbbqqf]

Hi! Triaging older issues — I think the API the OP described is largely shipped and this can probably be closed.

Most meta-learners (BaseSRegressor, BaseTRegressor, BaseXRegressor, BaseRRegressor, BaseDRRegressor) now expose estimate_ate(X, treatment, y) returning (ate, lb, ub) with the confidence-interval bounds derived from the bootstrap path (fit_predict(..., return_ci=True)) — see causalml/inference/meta/tlearner.py::estimate_ate and the corresponding lines in the other learners. P-values aren't returned in the same tuple but are derivable from the same bootstrap distribution; #890 also recently added rate_score() with bootstrap CI + p-values for the RATE metric.

If there's a specific learner / API surface that's still missing the SE+p-value tuple the OP asked for, a small follow-up issue pointing at the gap would be easier to action than this 2023 thread.

(Disclosure: I drafted this comment with help from Claude Code while triaging stale issues; the references above were verified manually against the current master branch.)