summary.causalmixgpd_causal_fit() returns posterior summaries for the
fitted PS block (when present) and both arm-specific outcome models.
# S3 method for class 'causalmixgpd_causal_fit'
summary(object, pars = NULL, ps_pars = NULL, probs = c(0.025, 0.5, 0.975), ...)
object:
A "causalmixgpd_causal_fit" object.
pars:
Optional character vector of outcome-model parameters to summarize in
both treatment arms. Passed to
summary.mixgpd_fit.
ps_pars:
Optional character vector of PS-model parameters to summarize. If
NULL, all monitored PS parameters are summarized.
probs:
Numeric vector of posterior quantiles to report.
…:
Unused.
An object of class "summary.causalmixgpd_causal_fit" with elements
ps, outcome, and probs.
This summary stays at the model-parameter level. It aggregates posterior summaries for the nuisance model \e(x)\ and for the arm-specific outcome models \f_0(y \mid x)\ and \f_1(y \mid x)\, but it does not yet collapse those pieces into treatment-effect functionals.
That separation is intentional. Parameters and treatment effects answer
different questions: summary.causalmixgpd_causal_fit() summarizes
posterior draws of the fitted model, whereas
ate(),
att(),
cate(),
qte(),
qtt(),
and
cqte()
transform those draws into causal contrasts.
print.causalmixgpd_causal_fit,
predict.causalmixgpd_causal_fit,
ate,
qte,
cate,
cqte.