fairlens.metrics.binominal_proportion_p_value#

binominal_proportion_p_value(p_obs, p_null, n, alternative='two-sided')[source]#

Calculate an exact p-value for an observed binomial proportion of a sample.

Parameters
  • p_obs (float) – Observed proportion of successes.

  • p_null (float) – Expected proportion of sucesses under null hypothesis.

  • n (int) – Sample size.

  • alternative (str, optional) – Indicates the alternative hypothesis. One of “two-sided”, “greater”, “less”. Defaults to “two-sided”.

Returns

The p-value under the null hypothesis.

Return type

float