anova

Likelihood-ratio test for comparing two nested linear mixed models. Mirrors lme4::anova.merMod() output format.

anova(m1, m2)[source]

Compare two nested linear mixed models via likelihood-ratio test.

Both models must have been fitted with method='ML'. When models differ in their fixed effects, REML-fitted models cannot be compared by LRT because the REML criterion depends on the fixed-effect structure.

Parameters:

m1, m2 – Two CrossedLMEResult objects. Order does not matter; the function sorts them by number of parameters (smaller model first).

Return type:

Any

Returns:

pandas.DataFrame –

Two-row table with columns matching lme4’s anova output:

Df  AIC  BIC  logLik  deviance  Chisq  Chi Df  Pr(>Chisq)

The first row corresponds to the simpler model; Chisq, Chi Df, and Pr(>Chisq) are NaN for it.

Raises:

ValueError – If either model was fitted with method='REML'.

Parameters:

Example

import interlace

# Both models must use method="ML"
m0 = interlace.fit("rt ~ 1",         data=df, groups=["subject", "item"], method="ML")
m1 = interlace.fit("rt ~ condition", data=df, groups=["subject", "item"], method="ML")

result = interlace.anova(m0, m1)
print(result)
#        Df       AIC       BIC       logLik  deviance  Chisq  Chi Df  Pr(>Chisq)
# Model 1  3  1234.56  1245.67  -614.28   1228.56
# Model 2  4  1228.12  1242.45  -610.06   1220.12  8.44       1    0.0037

Notes

Both models must have been fitted with method="ML". When comparing models that differ only in their random-effect structure, REML is acceptable and often preferred; use method="REML" in that case and pass both models with the same fixed effects.

See also