statsmodels.stats.oneway._fstat2effectsize(f_stat, df)[source]

Compute anova effect size from F-statistic

This might be combined with convert_effectsize_fsqu


Test statistic of an F-test

degrees of freedom df = (df1, df2) where
  • df1 : numerator degrees of freedom, number of constraints

  • df2 : denominator degrees of freedom, df_resid

resHolder instance

This instance contains effect size measures f2, eta2, omega2 and eps2 as attributes.


This uses the following definitions:

  • f2 = f_stat * df1 / df2

  • eta2 = f2 / (f2 + 1)

  • omega2 = (f2 - df1 / df2) / (f2 + 2)

  • eps2 = (f2 - df1 / df2) / (f2 + 1)

This differs from effect size measures in other function which define f2 = f_stat * df1 / nobs or an equivalent expression for power computation. The noncentrality index for the hypothesis test is in those cases given by nc = f_stat * df1.

Currently omega2 and eps2 are computed in two different ways. Those values agree for regular cases but can show different behavior in corner cases (e.g. zero division).