smmargins.DiDResult

class smmargins.DiDResult(cells: MarginsResult, simple_effects: MarginsResult, did: MarginsResult, joint: MarginsResult)

Bundle of results from Margins.did().

Holds the four cell predictions, the two simple effects, the difference-in-differences estimate, and a joint result that contains all three contrasts with their shared covariance.

Parameters:

• cells (MarginsResult) – The four adjusted predictions.

• simple_effects (MarginsResult) – Group effect evaluated at each level of the condition.

• did (MarginsResult) – The single difference-in-differences estimate.

• joint (MarginsResult) – All three contrasts (2 simple effects + DiD) sharing joint vcov, useful if you want to jointly test e.g. simple_effects = did = 0.

cells

Type:

MarginsResult

simple_effects

Type:

MarginsResult

did

Type:

MarginsResult

joint

Type:

MarginsResult

Examples

Use the bundle to grab whichever piece you want:

res = M.did("group", "preexist_Y",
            group_levels=["A", "B"], condition_levels=[0, 1])
print(res)                          # full report (cells + effects + DiD)
res.cells.summary()                 # 4-row table for plotting
res.simple_effects.estimate         # 2 group-effects, one per Y level
res.did.estimate                    # the single DiD on the response scale
res.joint.vcov                      # 3x3 covariance for joint Wald tests

Runnable smoke test:

>>> import numpy as np, pandas as pd, statsmodels.formula.api as smf
>>> from smmargins import Margins
>>> rng = np.random.default_rng(0)
>>> df = pd.DataFrame({
...     "treat": rng.integers(0, 2, 200),
...     "post":  rng.integers(0, 2, 200),
... })
>>> df["y"] = (df["treat"] + df["post"] + 0.5 * df["treat"] * df["post"]
...            + rng.standard_normal(200))
>>> res = Margins(smf.ols("y ~ treat * post", df).fit()).did("treat", "post")
>>> res.cells.estimate.shape
(4,)
>>> res.simple_effects.estimate.shape
(2,)
>>> res.did.estimate.shape
(1,)
>>> res.joint.vcov.shape
(3, 3)

__init__(cells: MarginsResult, simple_effects: MarginsResult, did: MarginsResult, joint: MarginsResult)

Methods

__init__(cells, simple_effects, did, joint)
summary()	Return the full DiD summary as one concatenated DataFrame.