statsmodels.stats.oneway.convert_effectsize_fsqu¶

statsmodels.stats.oneway.convert_effectsize_fsqu(f2=None, eta2=None)[source]¶

Convert squared effect sizes in f family

f2 is signal to noise ratio, var_explained / var_residual

eta2 is proportion of explained variance, var_explained / var_total

uses the relationship: f2 = eta2 / (1 - eta2)

Parameters:¶

f2 : None or float¶

Squared Cohen’s F effect size. If f2 is not None, then eta2 will be computed.

eta2 : None or float¶

Squared eta effect size. If f2 is None and eta2 is not None, then f2 is computed.

Returns:¶

res – An instance of the Holder class with f2 and eta2 as attributes.

Return type:¶

Holder instance