statsmodels.stats.diagnostic.het_goldfeldquandt

statsmodels.stats.diagnostic.het_goldfeldquandt(y, x, idx=None, split=None, drop=None, alternative='increasing', store=False, *, result_object=None)

Goldfeld-Quandt homoskedasticity test

This test examines whether the residual variance is the same in 2 subsamples.

Parameters:

yarray_like

endogenous variable

xarray_like

exogenous variable, regressors

idxint, default None

column index of variable according to which observations are sorted for the split

split{int, float}, default None

If an integer, this is the index at which sample is split. If a float in 0<split<1 then split is interpreted as fraction of the observations in the first sample. If None, uses nobs//2.

drop{int, float}, default None

If this is not None, then observations are dropped from the middle part of the sorted series. If 0<split<1 then split is interpreted as fraction of the number of observations to be dropped. Note: Currently, observations are dropped between split and split+drop, where split and drop are the indices (given by rounding if specified as fraction). The first sample is [0:split], the second sample is [split+drop:]

alternative{“increasing”, “decreasing”, “two-sided”}, default “increasing”

This specifies the alternative for the p-value calculation.

storebool, default False

Flag indicating to return the regression results

result_objectbool, optional

Flag indicating whether to return the results as a GoldfeldQuandtResult NamedTuple instead of a plain tuple. When store=True the NamedTuple holds the same four elements as the legacy tuple, so it unpacks identically and is always returned, with no warning. When store=False the legacy three-element tuple is returned by default and a FutureWarning is issued.

Deprecated since version 0.15.0: In release 0.16.0 or after July 2027, whichever is later, the default will change to always return a GoldfeldQuandtResult. Set result_object=True to opt in now, or result_object=False to silence the warning and keep the current return type.

Returns:

GoldfeldQuandtResult

If result_object=True, a NamedTuple with fields fval, pval, ordering, and res_store (res_store is None when not computed). See GoldfeldQuandtResult.

Otherwise (the deprecated default), a plain tuple made up of:

fvalfloat

value of the F-statistic

pvalfloat

p-value of the hypothesis that the variance in one subsample is larger than in the other subsample

orderingstr

The ordering used in the alternative.

res_storeResultsStore, optional

Storage for the intermediate and final results that are calculated

Notes

The Null hypothesis is that the variance in the two sub-samples are the same. The alternative hypothesis, can be increasing, i.e., the variance in the second sample is larger than in the first, or decreasing or two-sided.

Results are identical to R, but the drop option is defined differently. (sorting by idx not tested yet)