statsmodels.stats.nonparametric.cohensd2problarger

statsmodels.stats.nonparametric.cohensd2problarger(d)

Convert Cohen’s d effect size to stochastically-larger-probability.

This assumes observations are normally distributed.

Computed as

p = Prob(x1 > x2) = F(d / sqrt(2))

where F is cdf of normal distribution. Cohen’s d is defined as

d = (mean1 - mean2) / std

where std is the pooled within standard deviation.

Parameters:

dfloat or array_like

Cohen’s d effect size for difference mean1 - mean2.

Returns:

probfloat or ndarray

Prob(x1 > x2)