smmargins.MarginsResult

class smmargins.MarginsResult(estimate: ndarray, vcov: ndarray, labels: Sequence[str] | None = None, level: float = 0.95, df: int | None = None, stat_name: str = 'margin', outcome_labels: Sequence[str] | None = None, outcome_index: ndarray | None = None, ci_method: str = 'pointwise', draws: ndarray | None = None)

Container for marginal effects or adjusted predictions.

This class holds the point estimates, their delta-method covariance matrix, and provides methods for summarizing results, computing linear contrasts, or subsetting multi-outcome results.

estimate

Point estimates for each margin.

Type:

ndarray

vcov

The estimated covariance matrix of the margins (delta method or empirical covariance of simulation/bootstrap draws).

Type:

ndarray

se

Standard errors (square root of the diagonal of vcov, or the standard deviation of draws when available).

Type:

ndarray

tvalues

Test statistics (estimate / se).

Type:

ndarray

pvalues

Two-sided p-values based on either the normal or t distribution.

Type:

ndarray

labels

Labels for each row of the results.

Type:

list of str

ci_method

Method used to compute confidence intervals (“pointwise”, “bonferroni”, “sidak”, or “sup-t”).

Type:

str

draws

Simulation/bootstrap draw matrix of shape (S, m) when available.

Type:

ndarray or None

__init__(estimate: ndarray, vcov: ndarray, labels: Sequence[str] | None = None, level: float = 0.95, df: int | None = None, stat_name: str = 'margin', outcome_labels: Sequence[str] | None = None, outcome_index: ndarray | None = None, ci_method: str = 'pointwise', draws: ndarray | None = None)

Methods

__init__(estimate, vcov[, labels, level, ...])
contrast(c[, labels, name, ci_method])	Compute linear contrasts of the estimates.
outcome(k)	Slice rows belonging to the specified outcome class(es).
pairwise(by[, ci_method])	All pairwise comparisons of a factor variable's levels.
summary()	A summary table of the results.
wald([C, value])	Joint Wald test of linear restrictions on the margins.

Attributes

ci_lower	Lower confidence bound.
ci_upper	Upper confidence bound.
pvalues
se
tvalues